|
SOLID-STATE OLEFIN METATHESIS
By
GARRETT W. OAKLEY
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
2004
Copyright 2004
by
Garrett W. Oakley
This dissertation is dedicated to my family.
ACKNOWLEDGMENTS
First and foremost, I thank Jennifer Sippel/Oakley, the girl who became my wife on
October 2, 2004. I met her when I visited the University of Florida on my recruiting trip,
we started dating the day I moved to Florida, and we are getting married the same
semester that I plan to graduate. She has been there throughout my entire graduate
career, and the motivation, understanding, and love she has given me during the
completion of this work is irreplaceable.
I would like to thank my family, both immediate and extended, past and present, to
whom this document is dedicated, for being an endless source of support in my education
and life. Without this, I would have been lost. In particular, I want to thank my father,
David Oakley, for all his help and advice. Even after eight years of schooling and going
into debt, my father has encouraged me to get the best education I could. I thank my
mother, Jeanne Qualey, for the time in recent years that we were able to spend together
since she moved to Florida. This time has given me a renewed appreciation for the love a
mother can give. I also need to thank my mother for all the time she spent in raising me.
She, above all others, was the most instrumental in my development and the way I think
on a daily basis. I want to thank my brother, Brenton, and sister, Loren, for being my
family in the most loving sense of the word. I cannot imagine my life without them in it.
I thank my stepparents, Beverly Oakley and Thomas Qualey, for their love and support,
and for treating me like a son. I likewise thank my stepsister, Jackie, for treating me like
a brother, and for playing checkers every time I had a chance to visit. I would also like to
thank my grandparents, Wilmot and Ann Oakley and Cathy Garrett, as well as my late
grandfather, Roger Garrett, for the encouragement and love they have given me over the
years. Additionally, several aunts, uncles, and cousins deserve my thanks for their advice
and support.
I would like to acknowledge the people who helped mold my career choice of
chemistry, and then later the choice to specialize in polymer chemistry. I was first
exposed to chemistry in my sophomore year in high school by Mrs. Merle Cohen, who I
thank for being a wonderful teacher. The following year, I took AP chemistry from Mr.
Bruce Dinsmore. By the end of his class, I had decided that I wanted a career in
chemistry. I thank him for his logical teaching style, for preparing the class well for the
AP exam, and for telling us more than we needed to know, as I was more than ready to
take on sophomore organic chemistry as a freshman in college.
At Rensselaer Polytechnic Institute (RPI), I met Dr. Brian Benicewicz, whom I
thank for being the person most responsible for focusing my interest in chemistry towards
polymers. He was also the person who first suggested the University of Florida to me as
a place for graduate study. His enthusiasm for polymer chemistry truly inspired me, and
I thank him for being a very approachable research advisor and a friend. Other professors
at RPI that I would like to thank for their help and guidance include Dr. R. A. Bailey, Dr.
Tom Apple, Dr. Charles Gillies, Dr. Sonja Krause, Dr. Leonard Interrante, Dr. Wilfredo
Colon, and Dr. Kim Fortun. During my junior year at Rensselaer, I was fortunate to be
able to spend eight months on a co-op with the Bayer Corporation. I would like to thank
Dr. Ray Burk, Dr. Wan Wu, David Jenney, Dr. Inny Kim, Dr. Laura Spangler, and Tim
Rappl for a wonderful experience, which piqued my appetite for the polymer industry.
Ray and Laura deserve extra thanks for their encouragement in applying to graduate
schools.
At the University of Florida, I would like to thank all my coworkers on the George
and Josephine Butler Polymer Floor. Special thanks needs to be given to Stephen "Ed"
Lehman, Jr.; Barry C. Thompson; and Ben Reeves. Ed was a fellow RPI alumni and my
mentor in the laboratory, and I thank him for all the advice he has given me over the
years. Barry, my best friend and best man, has also been an invaluable resource of
knowledge in the lab, and has been a huge source of encouragement at every milestone in
achieving my degree. Barry, Ben, and I started graduate school together and supported
each other through classes, T-Hud, cumes, oral exams, and dissertation research, and I
thank them both for their help and support in all these obstacles. Other floor members to
whom I owe extra thanks for their help along the way are Jason Smith, Carl Gaupp, Jim
Pawlow, John Schwendeman, Pat O'Donnell, Tim Hopkins, John Sworen, Shane
Waybright, Avni Argun, Stephen Carino, Andrew Skolnik, Travis Baughman, Florence
Courchay, Piotr Matloka, Genay Jones, Christophe Grenier, Aubry Dyer, Emine Boz, and
Bob Brookins.
Enough thanks cannot be expressed to Lorraine Williams and Sara Klosser in the
polymer office. Without them, the little things that make everything run smoothly, but no
one wants to think about, would never get done. I also extend my thanks to Lori Clark
and Dr. Jim Deyrup for their help in the graduate office. I am extremely grateful to Dr.
George B. Butler and his wife Josephine for establishing the Butler Polymer Laboratory.
It has truly been a unique and wonderful place to work and study.
Many thanks are extended to my committee members, Dr. John R. Reynolds, Dr. J.
Eric Enholm, Dr. Danial R. Talham, and Dr. Anthony B. Brennan. Their efforts in
reading and discussing the material presented in this dissertation have been instrumental
in making it a stronger document.
Finally, I must express my sincerest appreciation to my research director and
supervisory committee chair, Dr. Kenneth B. Wagener. His door was always open and
whenever I talked with him, I always felt encouraged, with my motivation to work hard
revitalized. "Don't give up!" he would say, when things seemed the most bleak. Never
once have I seen him appear angry, as he always deals with situations in a calm and
collected manner. Dr. Wagener has given me a wealth of advice and guidance for things
both inside and outside the laboratory. I will forever be thankful for his presence in my
life.
TABLE OF CONTENTS
page
A C K N O W L E D G M E N T S ................................................................................................. iv
L IST O F T A B L E S ......... ........................................................ ............. ....... xi
L IST O F FIG U R E S .... ...... ...................... ........................ .. ....... .............. xii
A B S T R A C T .........x.................................... ....................... ................. xv
CHAPTER
1 INTRODUCTION TO SOLID-STATE POLYMERIZATION AND
O LEFIN M E TA TH E SIS .................................................... .............................. 1
1.1 Solid-State Reactions vs. Liquid-State Reactions ..............................................1
1.2 Crystal-to-Crystal Solid-State Polymerization ..................... .................. .......... 4
1.3 Solid-State Polycondensation of Semicrystalline Polymers..............................12
1.3.1 Monomer-to-Polymer Condensation SSP ..............................................13
1.3.2 Prepolymer-to-Polymer Condensation SSP ............................................15
1.4 O lefin M etathesis.......................................... .............. ........ ...... ... 23
1.4.1 Historical Overview of the Mechanism..........................................26
1.4.2 Development of Well-Defined Catalysts.................................................31
1.5 Opportunities of Olefin Metathesis in the Solid State................. ............ ...44
2 EXPERIM ENTAL M ETHODS ........................................ ........................... 47
2.1 The Solid-State A pparatus ...................................................... ............... 47
2.2 Making a Solid Mixture of Reactants .................... .................. ....................48
2.2.1 M elt Polym erization ............................... ............................................. 49
2.2.2 Freeze-Drying .......................................... .. .... ...... .............. 50
2.2.3 Grinding Together Diene and Catalyst...........................................51
2.2.4 Regrinding Clean Polymer with Fresh Catalyst............... ...................52
2.3 Sampling of the Solid Mixtures............................................ 52
2.4 Characterization Techniques Used on the Samples................... ................55
2.4.1 Gel Permeation Chromatography (GPC).................................................55
2.4.2 Nuclear Magnetic Resonance (NMR) Spectroscopy..............................56
3 SOLID-STATE ADMET OF SEMICRYSTALLINE POLYMERS .........................58
3 .1 Intro du action ............. ............. .......................... ................ 5 8
3.2 Experimental ....... ........ ............ ................ .. ...... .....62
3 .2 .1 G e n e ra l ................................................................................................. 6 2
3.2.2 M onomers Syntheses ......... ............... ..................... 63
3.2.2.1 Synthesis of tricosa-1,22-dien-12-ol (13).................... ............... 63
3.2.2.2 Synthesis of tricosa-l,22-dien-12-one (15)................................64
3.2.2.3 Synthesis of heneicosa-l,20-dien-11-one (16)..............................65
3.2.2.4 Synthesis of the polystyrene-graft macromonomer (19)..................66
3.2.3 Polym er Syntheses........................................................... ............... 67
3.2.4 Other M olecules Synthesized ........... ................ .................................. 67
3.2.4.1 Synthesis ofundec-10-en-l-ol (21)............................................68
3.2.4.2 Synthesis of 11-bromo-undec-l-ene (22)................. ............. ...68
3.2.4.3 Synthesis of 11-chloro-undec-l-ene (23)......................................69
3.3 R results and D discussion ............................................... .............................. 69
3.3.1 Prepolym er-to-Polym er Experim ents.................................... ..................69
3.3.1.1 Solid-state polymerization of 1,9-decadiene...............................69
3.3.1.2 Alcohol and ketone functionalized monomers.............................78
3.3.1.3 Polystryene-graft macromonomers ...............................................84
3.3.2 Monomer-to-Polymer Experiments......... .......................................85
3.3.3 Functional-Group Diffusion ..... ..................... ...............89
3.3.4 T herm al A naly sis........... ............................................ ........ .... .......... 9 1
3 .4 C o n c lu sio n s ..................................................................................................... 9 7
4 SOLID-STATE ADMET OF RIGID-ROD POLYMERS ......................................100
4 .1 In tro d u ctio n ................................................................................................... 1 0 0
4.2 Experim mental ............................................................. ......... 104
4 .2 .1 G en eral ...............................................................10 4
4.2.2 Monomer Syntheses ................................ ..................... 104
4.2.2.1 Synthesis of 2,5-dibromohydroquinone (25) ........... ................ 104
4.2.2.2 Synthesis of 1,4-dibromo-2,5-dipropoxybenzene (26) ................05
4.2.2.3 Synthesis of 2,5-dipropoxyterephthalaldehyde (27) ......................106
4.2.2.4 Synthesis of 1,4-dipropoxy-2,5-divinylbenzene (28)...................106
4.2.2.5 Synthesis of 1,4-dibromo-2,5-dioctadecyloxybenzene (29)...........107
4.2.2.6 Synthesis of 1,4-dioctadecyloxy-2,5-divinylbenzene (30)............. 108
4.2.3 Solid-State Polymerization Conditions ......................................... 109
4.3 Results and Discussion ............ .... ......... .... ........ 110
4.4 Conclusions .............. ...... .......... ............ ...................... 113
5 SOLID-STATE RING-CLOSING METATHESIS ...............................................114
5 .1 In tro d u ctio n ................................................................................................... 1 14
5.2 E xperim mental ............................. ............................................................. 117
5 .2 .1 G en eral .............. .. .. ......................................... 117
5.2.2 Synthesis of Dienes for Solid-State RCM .................................1...17
5.2.2.1 Synthesis of N,N-diallyl-3,5-dinitrobenzamide (31)...................... 117
5.2.2.2 Synthesis of 1-allyl-3-butenyl octadecanoate (32) .........................118
5.2.2.3 Synthesis of diallyl malonic acid (33)..................................... 119
5.2.2.4 Synthesis of the dipotassium salt of diallyl malonic acid (34)....... 119
5.2.3 Solid-State RCM Reactions Conditions................................................... 120
5.2.4 Synthesis of Ring-Closed Product (2,5-dihydropyrrol-1-yl)-(3,5-
dinitrophenyl)-m ethanone (35) .............. .......................... ................. 121
5.3 R results and D discussion ....................................................................... 121
5 .4 C on clu sion s................................................. ................ 12 7
LIST OF REFEREN CES ........................................................... .. ............... 129
BIOGRAPHICAL SKETCH ............................................................. ............... 140
LIST OF TABLES
Table page
1-1 Functional group tolerance of early and late transition metal olefin
m etathesis catalysts ...................... ...................... ..................... .. .... .. 34
1-2 Com prison of condensate boiling points ..................................... .................46
3-1 Results of solid-state polymerizations of monomer 11 with periodic
sa m p lin g .......................................................................... 7 0
3-2 Useful molecular weights of typical step-growth polymers..................................71
3-3 Results of solid-state polymerizations using Schrock's catalyst 2 and
m o n o m er 1 1 ...................................................... ................ 7 2
3-4 Catalyst loading comparison using monomer 11 and catalyst 6 ...........................73
3-5 Experiments conducted at 60 OC using monomer 11 ............................................75
3-6 Comparison of argon and vacuum experiments using monomer 11 and
c ata ly st 6 ......................................................................... 7 7
3-7 Examples of solid-state polymerizations of alcohol monomer 13 ...........................80
3-8 Examples of solid-state polymerizations of ketone monomer 16 ............................81
3-9 Precipitation effects on the solid-state polymerizations with catalyst 5
an d k eto n e 1 6 ...................................................................... 8 3
3-10 Results from the monomer-to-polymer freeze-drying experiment
u sing ketone 15 .......................................................................86
3-11 Results from the monomer-to-polymer experiments using ketone 16
and catalyst 6 (100:1 monomer to catalyst ratio)............................................. 87
3-12 Results of SSPs starting from a prepolymer of 18 without a tail on the
melting endotherm that extends into the reaction temperature .............................95
5-1 Results of solid-state RCM experiments...................................123
LIST OF FIGURES
Figure pge
1-1 Topochem ical solid-state polym erization ........................................ .....................5
1-2 The two types of topochemical SSP: (a) homogeneous growth starting
at randomly distributed point in the crystal; (b) heterogeneous growth
b y n u c le atio n ..............................................................................................................6
1-3 Photopolym erization of diacetylenes .............................................. ............... 8
1-4 Projection of the monomer and polymer crystal structures of the
bis(p-toluenesulfonate) diyne (TS)....................................... .......................... 9
1-5 Changes in unit cell parameters for PTS as the polymerization proceeds ...............9
1-6 Photopolymerization 2,5-distyrylpyrazine (DSP)................................................. 10
1-7 Conversion of (a) DSP into (b) dimer and (c) polymer..........................................11
1-8 Solid-state monomer-to-polymer polycondensation of 6-aminohexanoic
acid and 11-aminoundecanoic acid to Nylon 6 and Nylon 11, respectively ............14
1-9 Formation of a prepolymer of PET followed by an increase in the degree
of polymerization while in the solid state .................. ................... .............. 16
1-10 Reactive functional groups in the amorphous region that are anchored
in th e c ry sta l ....................................................... ................ 1 8
1-11 Typical plot of molecular weight versus time for a polycondensation SSP.............21
1-12 Transesterification interchange reaction taking place along the
polym er backbone ...................... ...................... .................... .. ......22
1-13 Modes of olefin metathesis in both organic and polymer chemistry .....................25
1-14 Early-proposed quasicyclobutane mechanism for olefin metathesis,
which was later found to be incorrect ........................................... ............... 27
1-15 Tetramethylene metallocycle mechanism for olefin metathesis, which
w as later found to be incorrect ...................................................................... ..... 28
1-16 Metal-carbene mechanism for olefin metathesis, which is now well
established as the correct m mechanism ................................... ........................ 30
1-17 Catalytic cycle for olefin metathesis in an example of RCM ...............................31
1-18 Common version of Schrock's tungsten (1) and molybdenum (2) based
olefin m etathesis catalysts ............................................. ..... ........................ 33
1-19 Select well-defined ruthenium metathesis catalyst ...............................................35
1-20 Associative and dissociative pathways of first- (5) and second- (6)
generation G rubbs catalyst ......................................................... .............. 39
1-21 Hoveyda's catalyst undergoing RCM, whereby the original
oxygen-coordinated catalyst is reformed ...................................... ............... 40
1-22 Bimolecular decomposition of catalyst 6 to form diruthenium
complex 10, which is likely responsible for isomerization during
o lefi n m etath esis........ ...................................................................... ...... ............ .. 4 3
1-23 Mechanism proposed by Grubbs for the decomposition of catalyst 6 ...................44
1-24 The ADMET mechanism whereby the equilibrium is shifted toward
the polymer by the release of ethylene ........................................... ...............45
2-1 Drawing of a typical apparatus for solid-state reactions..........................................48
2-2 Freeze-drying apparatus with liquid nitrogen trap .................................................51
2-3 Reaction of ruthenium catalysts with EVE ........................................ ......... .......53
2-4 Free radical scavenging with BHT................................................53
2-5 Relative NMR integration of internal and external olefins for NMR
end-group analy sis.......... .............................................................. .......... ....... 57
2-6 Olefin region in the NMR for typical ADMET polymers with long
methylene spacers. ................................................. ........ ...... 57
3-1 Monomer used in solid-state ADMET polymerization.................. ............... 61
3-2 Gel Permeation Chromatography traces of the SSP of 1,9-decadiene
with first-generation Grubbs at 45 C sampled periodically...............................71
3-3 Prepolymerization in the melt of 1,9-decadiene, followed by polymerization
in the solid state using second-generation Grubbs ruthenium catalyst (6)...............77
3-4 Synthetic route to alcohol 13 and ketone 15 .................................... ..................78
3-5 Mechanism of synthesis of ketone 16 ........ ........................ ......................... 79
3-6 Atom transfer radical polymerization (ATRP) of initiator 24 to
macromonomer 19 with subsequent ADMET to polymer 20 ................................85
3-7 The ADMET mechanism at work in the amorphous region of a
sem crystalline solid ......... ................... ................................ 90
3-8 Typical DSC showing the environment at the start of a
1,9-decadiene (11) experiment........ .. ......... .... ............ .............. ... 92
3-9 A DSC of a precipitated prepolymer from monomer 16......................................92
3-10 Both the DSC scans of the starting and 7-d samples from the
1,9-decadiene experiment in Figure 3-3..... ................... ...............93
3-11 A DCS trace of ketone polymer 18 that was synthesized in solution,
precipitated, and washed with boiling methanol to eliminate the tail
on the m elt endotherm .......... .................................. ................... 94
3-12 Typical GPC curve of polymer from monomer 11 and catalyst 6 ........................97
4-1 Indirect synthesis of PPV via a soluble polymer precursors
established by W essling ......................................................... .............. 102
4-2 Synthetic route to 1,4-dipropoxy-2,5-divinylbenzene (28)..................................110
4-3 Synthetic route to 1,4-dioctadecyloxy-2,5-divinylbenzene (30)............................111
4-4 Prepolymerization in the melt of 28, followed by polymerization in
the solid state using second-generation Grubbs ruthenium catalyst (6).................112
5-1 Comparison of the ADMET and RCM mechanisms ....................... ........1.15
5-2 Common RCM dienes found in the literature for testing features
of RCM (E = CO2Et) ........................................................................ 116
5-3 Diene synthesized for solid-state RCM experiments ...... .......................122
5-4 Ring-closing metathesis of amide diene 31 to compound 35 using
cataly st 5 ...........................................................................12 4
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
SOLID-STATE OLEFIN METATHESIS
By
Garrett W. Oakley
December 2004
Chair: Kenneth B. Wagener
Major Department: Chemistry
While organic reactions are most often performed in the liquid or melt state, the
investigation of solid-state organic conversions and solid-state polymerization (SSP) have
received varying degrees of attention for several decades now. Solid-state
polymerization is traditionally divided into two classes: crystal-to-crystal polymerization,
well established by Wegner; and equilibrium polycondensation of semicrystalline
polymers, which is widely used in industry. The principles governing these two types of
polymerization are quite different, for the former is based on topological chemistry, while
the latter is thought to occur in the amorphous region of semicrystalline polymers.
Olefin metathesis is an astonishing chemical phenomenon, whereby a single
chemical reaction has become widely used and studied in several, very different fields of
chemistry, in both academic and industrial settings. The principle mission of this work is
to demonstrate that a particular mode of olefin metathesis, acyclic diene metathesis
(ADMET), can operate in the solid state using the same principles used for
semicrystalline condensation polymers in industry. Experiments were aimed at
mimicking the industrial processes, including a prepolymerization period in the melt
state, extrusion of this prepolymer to a smaller particle size, and flow of an inert gas (or
vacuum) through the reaction vessel to aid in condensate removal. In most cases of
ADMET, the condensate that is released is ethylene, which gives it a distinct advantage
over polyester, polyamide, and polycarbonate industrial processes in that the condensate
is extremely volatile (BP = -104 C), and therefore high temperatures are not required for
condensate removal to drive the equilibrium toward high polymer.
The "benchmark" hydrocarbon monomer, 1,9-decadiene, was chosen for the initial
work because it lacks any polar or protic functionality that could react or complex with
the metathesis catalysts. Steady increase in molecular weight was observed in the solid
state, as low as room temperature, with the first- and second-generation Grubbs
ruthenium catalysts. Similar trends of molecular weight increase were observed with
monomers of varying functionality and architecture, which included ketone and alcohol
functionalized monomers and a polystyrene graft copolymer. Conclusions with potential
industrial implications were drawn from thermal analysis. We believe that low molecular
weight oligomer and unreacted monomer act as plasticizers for the promotion of solid-
state polymerization. In experiments where these species were removed, no significant
molecular weight increase was observed.
Additionally, solid-state ADMET of rigid-rod polymers and ring-closing metathesis
(RCM) were explored. Dialkoxy-divinylbenzene monomers were synthesized and used
in ADMET polymerizations, where slight increases of molecular weight in the solid state
were observed. Solid-state RCM to low conversions was also demonstrated with an
amide diene before a subsequent, visible, heterogeneous change, which we interpreted as
local melting point depression about sites of defects from the presence of RCM product.
CHAPTER 1
INTRODUCTION TO SOLID-STATE POLYMERIZATION AND OLEFIN
METATHESIS
1.1 Solid-State Reactions vs. Liquid-State Reactions
Most organic reactions are performed in the liquid or melt state because these states
possess the relevant molecular freedoms to allow the reaction to occur. A common view
among chemists, and even x-ray crystallographers, is that a molecular crystal is made up
of frozen chemical entities, which can only experience slight molecular motions. Indeed
this is true for many cases; however, a vast assortment of chemical reactions do occur in
the solid state and may have several advantages over performing the reactions in the
liquid state. These reactions have received varying degrees of attention in the literature
for several decades now.1'2 In some cases, a reaction may not even occur in the liquid
state, and the solid state provides the only route to the product.
When considering a reaction in the solid state, several factors not present in the
liquid state must be examined. In the case of a solid-state polymerization (SSP) for
example, some of these factors are the crystallinity of the monomer and polymer, the
polymorphic phase(s), nucleation and growth, sites of defects in the crystal, and particle
size. The one factor, however, that plays the most dominant role in the feasibility of any
solid-state reaction is the ability of the reacting groups to interact. Whether the solid
environment is crystalline, semicrystalline, or amorphous in nature, there must be inti-
mate contact between reacting groups in order for the reaction to occur.
This is extremely important when working within a crystalline environment, where
only very slight vibrational motions are allowed. For the reaction to occur, the two
reacting groups must be very close in proximity (-4 A) and must have the proper
orientation in order for the reaction to occur.3 It is an extremely fortuitous when a
reaction of this nature can proceed in a crystalline state because the crystal structure of a
molecule is not something that is chosen by the scientist. Mother Nature decides the best
way for a particular molecule to stack together, and that structure will never change for
that particular molecule under a standard set of conditions. It is possible to change the
crystal structure by changing the crystallization procedure, although this is quite rare.
Usually, to change the crystalline structure, the molecular structure must change, thereby
changing the molecule being dealt with. This is not always a flexibility that can be
afforded. Changing the chemical structure of the molecule often changes the properties
of the product as well. This may be very limiting depending on the desired application.
Tailoring of a crystal structure to suit the needs of solid-state reaction is also impractical,
since space-filling patterns are extremely difficult to predict. A slight change at one site
on a molecule can have huge, unexpected changes in crystal structure, which could
change its space-filling pattern, and even its lattice type.
For most organic reactions where the starting material is a solid, the crystal
structure is just not conducive to performing the reaction in the solid state. In addition,
multiple reagents often are needed for the reaction to occur. Every chemist knows that
for most reactions, simply mixing two solids together will result in no reaction. Solvents
or melting are needed to allow the reacting species the intimate contact they need for the
reaction to occur by offering the much higher degree of molecular motion allowed in a
liquid state.
Amorphous or partially amorphous environments offer elevated mobility for
functional groups to meet. Above the glass transition temperature (Tg) of a polymer, the
amorphous areas are softened from their glassy state. Chains within the amorphous
region may shift their position, increasing the probability of reactive groups meeting,
even if the chain is anchored in a crystalline region.
There are several examples where solid-state polymerization does occur.3'4'5 The
main interest in studying SSP is to obtain polymers that are unattainable in the liquid
state or polymers that have better properties than those obtained in the liquid state. One
of the most desirable features of SSP is that some polymers have been shown to go to a
higher extent of reaction and achieve higher molecular weights than those obtained in the
liquid state. It is well known that many polymeric properties can be improved with
increase in molecular weight and SSP can provide a route to them. Solid-state
polymerization may also provide a route to intractable polymers (i.e., polymers that are
insoluble in almost every solvent and precipitate out of solution even at the dimer or
trimer stages of the polymerization). Special polymer morphologies are possible because
the reactions are occurring in the solid state. In many cases, the polymer created has a
high degree of crystallinity and/or stereoregularity that would be difficult or impossible to
achieve by any other means. Unwanted side reactions that lead to undesirable by-
products are limited or avoided because of close packing in the crystal. It has been
shown that polymers prepared by SSP often have less cyclic oligomer than those
prepared by other methods.6 Formation of crystals with fewer defects, such as spurious
moieties in the chains or chain branching, has also been demonstrated.7 Another reason
for these observations, aside from being in the solid state, is that the reaction is done at
lower temperatures than in the melt or solution. The high temperatures needed to keep
the material in the melt or in solution are not needed, so fewer side reactions occur. In
addition, SSP is commercially appealing because no solvent is required, there is
practically no pollution, and it requires less energy that it would to keep the polymer in
the melt state.
Solid-state polymerization is traditionally divided into two classes:8 crystal-to-
crystal polymerization,3'9 well established by Wegner; and equilibrium polycondensation
of semicrystalline polymers,4'5 which is widely used in industry. The principles
governing these two types of polymerization are quite different, for the former is based
on topological chemistry, while the latter is thought to occur in the amorphous region of
semicrystalline polymers.
1.2 Crystal-to-Crystal Solid-State Polymerization
Crystal-to-crystal solid-state polymerization actually occurs in a crystalline state.
Therefore, it relies on a suitable arrangement of the reactants in the monomer crystal to
allow for the necessary contact between reacting groups with only the very slight
vibrations allowed in a solid crystal. These reactions occur by only very slight rotations
of the reactants in a crystal (Figure 1-1). The ideal case is a single-crystal-to-single-
crystal transformation of monomer to polymer. The only example of SSP where this is
actually known to occur is the polymerization of certain diacetylenes in the solid
state.9,10,11
Figure 1-1. Topochemical solid-state polymerization
Most topological SSPs occur with a higher degree of disorder including examples
with nucleation of the polymer from a site of disorder or diffusion of reactants to centers
of reactivity. This can lead to complete destruction of the parent crystal with a high
degree of irregularity in the daughter crystal. However, the interesting examples retain
some degree of order in the daughter crystal. If a polymer crystal obtained in the solid
state retains a small number of crystallographically equivalent orientations of the mono-
mer (usually in one direction determined by the parent crystal structure) it is termed
topotactic. In contrast, a reaction is termed topochemical if the reaction is governed by
the geometry and spatial symmetry of the adjacent molecules and the reaction pathway is
uniquely defined by the parent crystal structure. Topochemical polymerization can be
thought of as a diffusionless transformation of monomer single crystal to polymer single
crystal. All topochemical reactions are topotactic, but there are examples of topotactic
reactions that are not considered topochemical. One such example is ring-opening
polymerization of oxacyclic compounds, such as trioxane and tetraoxane, which occurs
with diffusion of monomer. These types of examples that have a large degree of disorder
in the parent crystal will not be discussed further, but have been reviewed in the
literature. 3,12
Topochemical reactions can be either homogeneous or heterogeneous. In a homo-
geneous polymerization, chain growth starts at randomly distributed points throughout
the crystal lattice, and then forms extended chains of polymer in a solid solution (Figure
1-2a). Wegner,9 a pioneer in the field of the only known homogeneous SSP, describes it
as "a solid solution of individual extended chains of varying length all stretched out along
a unique crystallographic direction inside the matrix of the as yet unreacted monomer."
Heterogeneous solid-state polymerizations (Figure 1-2b) are the more common type.
I 'I ii- -
_(a) -(b)
Figure 1-2. The two types of topochemical SSP: (a) homogeneous growth starting at
randomly distributed point in the crystal; (b) heterogeneous growth by
nucleation
For the heterogeneous type, the reaction preferentially starts at defect sites in the
parent crystal, and proceeds with nucleation in the daughter phase. This process of
nucleation and phase growth results in fragmentation of the parent crystal, with a higher
probability of nucleation at the surface of an already existing nucleus. The formation of a
nucleus of an amorphous polymer inside the monomer crystal undoubtedly will cause
stresses within the crystal. These stresses are relieved by the formation of dislocations
around the original nucleus, which produces a new generation of reaction sites for
nucleation.' This in turn will produce a multitude of small crystallites in a solid, as
opposed to a large single crystal.
It happens that the best examples for both homogeneous and heterogeneous SSP
are polymerizations through unsaturated bonds. The homogeneous example is that of the
diacetylenes,9'10'11 which polymerize via 1,4-addition of conjugated triple bonds; and the
heterogeneous example is that of diolefins,13'14 which undergo [2+2] cycloaddition to
form cyclobutane rings in the backbone of the polymer. For a truly topochemical crystal-
to-crystal transformation, the unsaturated bonds in both cases must be mutually oriented
in the crystal. In this mutual orientation of the molecules, the unsaturated bonds of
adjacent molecules are close to parallel, preferably coplanar to each other, and not
exceeding a distance of 4 A between them.3
The polymerization of conjugated diacetylenes in the solid state occurs thermally or
with high-energy photoradiation. The polymerization occurs by a 1,4-addition with
diradical and dicarbene intermediates to give rise to a fully conjugated main chain
(Figure 1-3), where the polymer produced has two resonance structures. One structure
has three cumulated double bonds per repeat unit, while the other has alternating double
and triple bonds in conjugation. The diacetylene rods undergo slight rotational
movement with subsequent addition to the nearest neighbor during the polymerization.
X-Ray data has shown that in a crystal of polymeric diacetylene, the substituents adopt an
all-trans configuration.15
The crystal packing of the monomer is determined by the kind of substituents on
the diacetylene. These polymerizations can only proceed in the solid state if the
substituent groups interact strongly with intermolecular hydrogen-bond formation or
dipolar forces. The mutual position of the triple bonds in the crystal structure, and
therefore the reactivity of the monomer, is controlled by how closely the substituents
allow the molecules to pack together. If the intermolecular forces are not strong enough
to create a closely packed structure, then the polymerization will not proceed. A typical
example of a diyne that undergoes a successful SSP is the bis-p-toluenesulfonate of
2,4-hexadiyne-1,6-diol, abbreviated TS.16 The polymer product is thus abbreviated PTS.
R
RR R R
R
R R R R R
Figure 1-3. Photopolymerization of diacetylenes
Figure 1-4 shows a comparison of the atomic positions of the monomer and
polymer of PTS. Note how there is only a very slight change in the crystal structure
during the course of the reaction. This allows the polymerization to proceed without any
appreciable change in the original shape and size of the monomer crystal. Figure 1-5
illustrates the very small changes in unit cell parameters during the course of the reaction.
Molecular weights of PTS reach orders of -106 g/mol. At 100 % conversion, the
distance between chain ends is very small (less than 3 A); therefore, the polymer chains
can be considered as continuous lines extending from one end of a crystal to the other.
Consequently, the product of a polymerization of a diacetylene occurring in this manner
is considered a polymeric single crystal. This is the only polymer chemically occurring
in the form of a macroscopic single crystal; hence it is the only known homogenous SSP.
O O
O O
--//S-OCH, CH20-S-- \ TS
o0 0
b
450
Monomer { 5.11 A
Polymer 4.91 A
14.5
Figure 1-4. Projection of the monomer and polymer crystal structures of the bis(p-
toluenesulfonate) diyne (TS)
%X.X. X.XXX-XX X--X X-X 15.0
xxx
14.6
5.2 144
5 1 -
0 20 40 60 80 100
Polymer content (%)
Figure 1-5. Changes in unit cell parameters for PTS as the polymerization proceeds
A typical example of a heterogeneous SSP is the [2+2] photopolymerization of
diolefines.13'14 In contrast with the topochemical polymerization of diacetylenes, this
reaction proceeds by a step-growth mechanism. Upon UV radiation of wavelengths of
450-350 nm, double bonds on adjacent molecules undergo a [2+2] photocycloaddition to
form cyclobutane rings within the polymer backbone. The monomer does not necessarily
move much during the polymerization. However, this slight movement causes the
reaction to proceed by phase separation with nuclei growth that can advance in different
crystallographic directions at different speeds. This is opposed to chain growth in one
crystallographic direction, as in the case of a homogeneous diacetylene solid-state
polymerization. Although the result is destruction of the parent crystal, separate well-
defined phases can be observed in the crystal even in the early stages of the reaction.17
The first example of this four-center type photopolymerization of a crystal to be
studied was the polymerization of 2,5-distyrylpyrazine (DSP)18 (Figure 1-6). The result
of the polymerization is a highly crystalline linear polymer produced in quantitative
yield.
N hv N-
N [2+2] N
DSP
S-
poly-DSP
Figure 1-6. Photopolymerization 2,5-distyrylpyrazine (DSP)
The change in crystal structure as the polymerization proceeds is illustrated by
Figure 1-7. The distance between the reacting olefins in the parent crystal has been
found to be 3.939 A,19 which is within the 4 A criteria discussed earlier for a truly
topochemical reaction to occur.
(a) ` (b) Y (c)
Figure 1-7. Conversion of (a) DSP into (b) dimer and (c) polymer
In some other cases of diolefins that could potentially undergo the [2+2] cyclo-
addition, the propagation is hindered by an increase in the distance between reaction
sites, even after dimer is formed, due to the molecular displacement of the reaction. The
double bonds become too far apart for the reaction to propagate further.
The symmetry of the molecular arrangement is preserved during the course of the
polymerization of DSP, as indicated by the coincidence of the space groups (Pbca) of the
monomer and the polymer crystals. The reaction is therefore termed a lattice-duplicating
polymerization. The only significant change in the crystal structure is a 1.8 % con-
traction in the c direction from monomer to polymer.12
Of the several other types of heterogeneous SSP, a few are worth mentioning.
Polymerization of 1,3-dienes via a 1,4-addition in the solid state has recently been
reviewed.20 Some vinyl monomers can undergo a radical chain mechanism in the solid
state with a high degree of disorder. Lastly, there are even some examples of SSP of the
polycondensation type that transform from crystal-to-crystal in the solid state. One such
example is the formation of polyesters through thermally induced polymerization of
halogenoacetates.5
1.3 Solid-State Polycondensation of Semicrystalline Polymers
Solid-state polycondensation of equilibrium step-growth polymers was first
observed more than 40 years ago.21 22'23'24 This type of SSP is distinctly different from
most topochemical SSPs. One clear difference is that a small molecule is released from
the reaction for every propagation step that occurs. Since many polycondensation
reactions are reversible, this condensate must be removed in order to drive the
equilibrium toward the polymer and achieve high molecular weight. The diffusion of this
side product to the surface of the particle can become the rate-limiting factor if there is
not sufficient surface area to allow the condensate to escape the solid. The condensate is
most frequently removed by static or dynamic vacuum or by a stream of an inert gas.
Recently, it has been shown that supercritical CO2 may also be used as a sweep
fluid.25,26,27,28
Another distinct difference is that these are usually semicrystalline polymers, where
the polymerization is conducted above the glass transition temperature (Tg), but below the
melting point (Tm) of the polymer. This greatly enhances molecular mobility within the
amorphous region of the polymer, which facilitates end-group encounters and aids in
condensate diffusion. Since the reaction is conducted below the semicrystalline melting
point, the crystalline regions remain intact, and the particle maintains its shape during the
polymerization.
There are two distinct starting points for polycondensation reactions in the solid
state. The atypical case starts with a crystalline monomer, and goes all the way to the
polymer in the solid state. This is a rare occurrence, but it does potentially have great
scientific and commercial interest. More frequently however, the starting material has
already been polymerized to some extent in the melt or solution state, and is further
polymerized as a solid, leading to higher molecular weight.
To term this type of polymerization as a "solid-state polymerization" causes
discontent among some scientists. For one, the reaction is occurring above the glass
transition temperature of semicrystalline polymers, so there is a higher amount of
mobility in the amorphous region than in a crystalline solid, which many scientists
consider to be the "solid-state." Additionally, after the prepolymerization in the melt,
there may only be few couplings per chain needed in the "solid-state" portion of the
process to achieve the desired, higher molecular weight. The word, "polymerization," by
definition, means that many couplings are taking place in the process. However, to not
call this a "solid-state polymerization" will contradict more than 40 years of literature and
industrial terminology. A more suitable definition for this type of SSP is a step-growth
polymerization of a semicrystalline polymer, where the crystalline region maintains the
shape of particle while the polymer is growing.
1.3.1 Monomer-to-Polymer Condensation SSP
Before a thorough consideration of the solid-state polymerization of prepolymers,
further mention of monomer-to-polymer polycondensation is worthwhile. These exam-
ples span the two classes of SSP: crystal-to-crystal and polycondensation. Like any
crystal-to-crystal SSP, reactivity and the kind of crystalline order in the product are
determined to a considerable degree by the characteristics of the parent crystal. Again,
the parent crystal for monomer must have an acceptable arrangement of molecules, in
which the reacting functional groups are in close enough proximity and have the proper
orientation to allow the reaction to occur in the solid state. It is remarkable, however,
that this occurs despite the need to provide for the diffusion of low molecular weight
molecules to the surface of the particle. Reductions in volume and contractions of a
crystal along the direction of the chain have been observed during some SSPs from the
elimination of the condensate. One such example is the polymerization of 6-amino-
hexanoic acid to Nylon 6 (Figure 1-8).29 It was shown to have an 18 % reduction in
volume and a 17 % contraction in the direction of the polymer chain. Polymerization of
11-aminoundecanoic acid to Nylon 11 (Figure 1-8) was found to have a similar effect
with 16 % reduction in volume and 11 % contraction.4
160-170 C H
H2N OH-- vacu N + n H20
OH vacuum
x x
x = 1 6-Aminohexanoic acid x= 1 Nylon 6
5 11-Aminoundecanoic acid 5 Nylon 11
Figure 1-8. Solid-state monomer-to-polymer polycondensation of 6-aminohexanoic acid
and 11-aminoundecanoic acid to Nylon 6 and Nylon 11, respectively
Some monomers have been shown to polymerize to some extent in the solid state,
but not to the same degree of polymerization that can be achieved in the melt. An
example of this is polymerization of bis(2-hydroxyethyl terephthalate) to produce
poly(ethylene terephthalate) (PET; Figure 1-9). An example of this is polymerization of
bis(2-hydroxyethyl terephthalate) to produce poly(ethylene terephthalate) (PET; Figure
1-9). Poly(ethylene terephthalate) is a typical example of polycondensation SSP by
which a prepolymer of low molecular weight can achieve a higher degree of polym-
erization in the solid state, 30 as discussed next. However, if starting a solid-state
polymerization from a crystal of the monomer, it is very slow indeed, and it is virtually
impossible to achieve high molecular weight starting from monomer. Bamford and
Wayne30 reported dimer as the primary product, both with and without catalyst, based on
melting-point correlations.
1.3.2 Prepolymer-to-Polymer Condensation SSP
Solid-state polycondensation of semicrystalline polymers is widely practiced in the
polyester, polyamide, and polycarbonate industries. Typically, the polymerization begins
in the liquid (melt) state at high temperatures (usually above 250 C) to achieve a
moderate molecular weight, and then the polymerization is continued in the solid state to
achieve higher molecular weight. High molecular weights, often unattainable in the
molten state, can be obtained by SSP of prepolymers, leading to desirable physical
properties. Nylon 6,6 has been shown to achieve molecular weights of 280,000 g/mol
from starting material of 40,000 g/mol; a 7-fold increase in molecular weight while in the
solid-state.31
For the polyester, poly(ethylene terephthalate) (PET), the monomer is synthesized
by reaction of the terephthalic acid or its dimethyl ester with ethylene glycol to generate
the bis(2-hydroxyethyl) ester. This is done to alleviate the need for a stoichiometric
balance of two monomers, as is the case in most step polymerizations. The chemistry is a
simple transesterification that is reversible if the by-product, ethylene glycol, is not
removed.
Figure 1-9 shows the process by which this polymer is synthesized. First, the
monomer is polymerized in the melt. The molecular weight achieved during this first
process is approximately 15,000 g/mol, which is high enough for use in most
applications. This polymer is then cooled to a temperature below its semicrystalline
melting point, and extruded into small particles to increase its surface area. The polymer
chemistry is then continued in the solid state above the glass transition temperature, but
below the melting point of the semicrystalline polymer (usually about 150 C) to achieve
higher molecular weight. A molecular weight of approximately 25,000 g/mol can be
achieved for PET in the solid state, which can be used in more demanding applications.
O O ~-280 oC
S+ HO OH
HO 0 0 OH Vacuum --0 O
(melt) V
Prepolymer
0 O0
--0 0
x
Tg < Temp < Tm
Nitrogen Flow
(solid state)
+ HO OH
y >>x
Figure 1-9. Formation of a prepolymer of PET followed by an increase in the degree of
polymerization while in the solid state
Completing the polymerization at a lower temperature has distinct advantages over
performing the polymerization completely in the melt state. It is economically practical
because less energy is required to heat the reaction mixture, and because polymers of
higher purity are produced due to fewer side reactions occurring, and this can, in turn,
lead to higher molecular weights and properties unattainable in the bulk state.
As in many engineering plastics, molecular weight plays an important role in
achieving the desired properties of the polymer. Certain processes, such as blow-molding
and extrusion, require that the molten polymer has a high enough melt viscosity to
prevent collapse in the soft preformed state. This requires high molecular weights, which
can be obtained only with great difficulty in conventional melt processes. To achieve this
in the melt, extremely high temperatures are needed to lower the viscosity in order for the
reaction to occur, and therefore side reactions occur that counteract the desired increase
in molecular weight. In addition, polymer melts with high melt viscosities are difficult to
handle and hard to remove from bulk polymerization equipment. These disadvantages
can be simply avoided if SSP is an option.
Solid-state polymerization of semicrystalline prepolymers is a complex process.
The prepolymer is extruded into small chips or a powder. Tumble dryers or moving bed
reactors are usually employed to keep the small particles from sticking together. The
polymerization is continued at a reaction temperature well above the glass transition
temperature of the polymer, but below its melting point (Tg < Temp < Tm). A typical SSP
takes place 5-40 C lower than the melting point of the polymer.4 The small particles are
then subjected to a stream of inert gas or vacuum to remove the condensate. Work has
been done comparing the two;32 however, the choice is usually based on convenience.
Removing the condensate at the surface of the particle drives the equilibrium toward a
higher molecular weight polymer, improving the physical properties.
Usually, the shape of the material is maintained during a solid-state polymerization,
and then the product polymer can be extruded into its desired shape. If the polymer is
molded into its final form after the polymerization in the solid state, it is termed a
postextrusion SSP process. This is the only option unless the desired sculpt of the
product has at least one dimension small enough to allow the condensate to escape. Then
a preextrusion SSP process is possible, in which the prepolymer is fashioned into its
desired shape first, and then further polymerized in the solid state while maintaining the
same shape. Such can be the case in thin films and fibers.
Several factors play a role in the overall reaction kinetics of a solid-state
polycondensation. These include the crystallinity of the polymer, particle size, initial
molecular weight, number and type of chain end groups, reaction temperature, time of
polymerization, diffusion of condensate to the surface of the particle, and technique used
to remove the condensate at the surface. Most reports in the literature only treat a few of
these at a time, mainly treating the factors that affect the overall rate of the process.
It is generally accepted, but not proven, that the polymerization occurs in the
amorphous region. The rate of reaction is therefore restricted by the molecular mobility
of the reactive end groups, which are anchored down in the crystalline phase shown by
Figure 1-10.
Crystal
Functional Functional
I I Crystal
Figure 1-10. Reactive functional groups in the amorphous region that are anchored in the
crystal
In the kinetic treatment of data, it is sometimes assumed that all the end groups are
in the amorphous region to explain an increase of crystallinity with increased molecular
weight. This is supported by a study on the degree of crystallinity of polyamides at a
constant pressure.33 If this assumption is correct, the concentration of end groups in the
amorphous region would be higher than if they were distributed throughout the entire
volume of the solid, and the reaction would be shifted towards the polymer in order to
reestablish equilibrium. The result is an increase in molecular weight.
The degree of crystallinity may also have an effect on the diffusion coefficients of
the condensate.24'34 The higher the degree of crystallinity, the harder it is for the
condensate to diffuse through the solid. This is coupled with the fact that the degree of
crystallinity of some polymers increases as the solid-state polymerization proceeds.34
Newly formed chain extensions at the surface of the crystalline phase can be taken into
the crystal, increasing the percent crystallinity. This phenomenon is called SSP-induced
crystallization. The surface of larger particles will become more crystalline than the
inner portions of the particle since the condensate can more easily escape. This, in turn,
makes it even more difficult for the condensate to diffuse to the surface and escape from
the inner portions of large particles.
The particle size is the most unique factor to play a role in polycondensation SSP.
In general, while the overall rate increases with a decrease in particle size,30 the kinetics
can be much more complicated. With extruded geometries that allow for ease of escape
of the condensate in at least one dimension, diffusion of this side product to the surface
becomes a negligible part of the picture. The condensate does not have an opportunity to
undergo depolymerization if it does not have the opportunity to collide with a polymer
molecule. At larger sizes, other factors, such as temperature, play a larger role.
Poly(ethylene terephthalate), for example, at sizes greater than 100 mesh is diffusion
controlled at 210 OC. However, at 160 OC, the chemical reaction is slower, and the
chemical kinetics becomes the rate-limiting factor. This suggests that at temperatures in
the range of 160-210 C, the kinetics are controlled by both diffusion and chemical
kinetics.24
At particle sizes where the reaction is diffusion controlled, sections of a polymer
chip have been shown to have decreasing intrinsic viscosity as penetration into the center
of the polymer chip increases.30 This is evidence that the reaction is occurring primarily
at the surface of the particle since the higher molecular weight portions are more viscous.
Broadening of the overall molecular weight distributions have also been observed with
increased particle size.35 Larger particles will have a broader molecular weight gradient
in the polymer chip, causing the overall molecular weight distribution of the sample to be
broader.
Choice of temperature is one of the most important factors because it affects almost
all other aspects of the reaction. Too high a temperature may lead to partial melting with
concurrent sticking of the polymer chips or a change in the nature of the process, while
too low temperatures may prevent chain growth or lead to long reaction times. The
temperature for most SSPs is very high due to the high activation energy of the overall
process. Typically, the temperature of the SSP is 5-40 C lower than the melting point of
the polymer to overcome this activation energy.4 The lowest temperatures where SSP
takes place in a detectable manner are approximately 150-170 C for polyesters and
90-120 C for polyamides.
The time of the reaction is another important factor. A typical plot for a SSP of the
molecular weight vs. time is shown in Figure 1-11. The molecular weight reaches a
maximum and eventually drops down due to the occurrence of degradation reactions.
Optimization of reaction time must be done to achieve the most desirable molecular
weights in a SSP.
Mol. Wt
A
Time
Figure 1-11. Typical plot of molecular weight versus time for a polycondensation SSP
As in any reaction, much is dependent on the nature of the reagents used. The
initial molecular weight and the number and type of end groups of the prepolymer
determine what kind of molecular species will be involved in the reaction. The mobility
of the catalyst employed can also be a factor. Titanium catalysts have been used in
polyester synthesis as a coordination catalyst.36 It might be that the larger the amount of
catalyst in the mixture, the less diffusion of the catalyst is required for it to be effective.
Much of the discussion regarding solid-state polycondensation thus far has been
centered on the diffusion of some species through a solid. In addition to the diffusion of
small molecules to the surface, the migration of end-group functionality also needs to be
considered. Most models that consider the migration of end groups rely on the diffusion
of segments of polymer chains within the solid in order for end groups to meet and react.
However, rapid buildup of molecular weight cannot be rationalized by diffusion of
molecules or their segments alone, since migration of end groups would be required for
frequent encounters of reactive groups.31 Thermal motions are severely restricted in the
solid state, and segmental diffusion is considerably restricted as compared to the melt
state. In addition, the motions of most chain segments in semicrystalline polymers are
restrained by the fact that they are anchored in the crystal. A more recent model suggests
that end group migration can be explained by interchange reactions. Although not
recognized by most as the dominant mechanism, this model has received some
acknowledgment in recent years.31,37
In a polycondensation reaction, if the end group reacts with another end group, then
a small molecule is released. However, an end group also has the ability to react with
any functional groups along the chain backbone in an interchange reaction, such as the
transesterification shown in Figure 1-12. The result is just an exchange of functionality
and migration of a chain end with no affect on the number average molecular weight.
However, these exchanges of functionality facilitate the meeting of two end groups,
which can, upon reaction, release a small molecule and drive the reaction toward a higher
molecular weight.
OH R/ ROH OR O'R
SO OO
Figure 1-12. Transesterification interchange reaction taking place along the polymer
backbone
There is significant indirect evidence that these exchange reactions occur during a
SSP. Lenz and coworkers38'39 studied reorganization of copolymers, mostly copolyesters,
which showed an increase in the extent of blockiness of the copolymer upon heating in
the solid state. This requires that interchange reactions are taking place. A semicrys-
talline copolymer can undergo a reaction-facilitated crystallization and develop an
increasingly blocky structure if its repeat units can undergo interchange reactions. The
newly interchanged units at the surface of the crystalline interface are taken into the
crystalline region and thus become inaccessible to further reaction. These results also
show that interchange reactions are confined to the amorphous region of the polymer,
being severely restricted in the crystalline region.
In the literature, a small number of papers explicitly recognize interchange
reactions as the dominant mechanism for migration of functionality of polyesters and
polyamides.31'37 Interchange can result in the formation and swapping of loops and
bridges between crystalline regions. Abhiraman and coworkers31 predict that the
concentration of loops and bridges, such as those in Figure 1-12, will equal at equilibrium
if exchange reactions are taking place. They also found that the jump distance between
an amine and an amide in a SSP of Nylon 6,6 was found to be 5.12 A, which is in
excellent agreement with the 5-6 A, which has been obtained as the critical reaction
distance from classical studies of diffusion-controlled reactions of small molecules. This
means that the model can be thought of as being analogous to a diffusion-controlled
bimolecular system with one or both species diffusing through the system, depending on
whether or not one or both functional groups in the polymer can undergo exchange
reactions. This model seems to be gaining more and more support as time passes.
1.4 Olefin Metathesis
Throughout the history of chemistry, any reaction that has the ability to form
carbon-carbon bonds receives a significant amount of attention; olefin metathesis is no
exception. Olefin metathesis is an astonishing chemical phenomenon, whereby a single
chemical reaction has become widely used and studied in several, very different fields of
chemistry in both academic and industrial settings. The diversity in application of this
reaction has lead to its prevalence in the fields of organometallics,40 organic synthesis,41
and polymer chemistry,42 which have all been thoroughly reviewed.43'44'45 The word,
metathesiss," derived from the Greek words meta (change) and tithemi (place), means an
exchange; thus the term olefinn metathesis," originally coined by Calderon in 1967,46
refers to the interchange of carbon atoms (with their substituents) between a pair of
alkene bonds.44 Modern olefin metathesis offers a mild method of forming carbon-
carbon double bonds. In addition, the double bond offers functionality, which can be
subsequently used in a number of other chemistries.
There are several distinct modes under which olefin metathesis can operate. These
are illustrated in Figure 1-13. Cross metathesis (CM)47'48'49 is olefin metathesis in its
most literal sense, whereby two olefins are exchanging carbon atoms to make new
olefins. Here, the selectivity of the reaction is of the utmost importance since a
competing reaction is a self-metathesis of an olefin with itself to give the cross product.
Ring-closing metathesis (RCM)50'51'52'53 is just the intramolecular version of CM, by
which dienes can close to form cyclic structures with the release of ethylene. This
condensation reaction has found wide use in the synthesis of sophisticated organic
molecules. Chiral catalysts have been employed for asymmetric RCM, in which one
enantiomer of a racemic diene is ring-closed or an achiral diene is ring-closed chiro-
selectively.53'54 Ring-opening metathesis (ROM)55 is simply the reverse of RCM,
whereby an acyclic olefin is introduced to the cyclic olefin to yield acyclic dienes as
products. Like RCM, this process can also be conducted asymmetrically whereby the
added acyclic olefin preferentially adds to one position on the cyclic olefin.55
CM +
R1 R2 R1
+
\ O + -
Self- +
R2 Self- R R1
Metathesis
RCM
ROM
R RROM
6 ROMP
L ^j n
R ADMER + =
n
Figure 1-13. Modes of olefin metathesis in both organic and polymer chemistry
There are two modes of polymerization for olefin metathesis that are governed by
very different principles. Ring-opening metathesis polymerization (ROMP)42'56'57'58 is a
chain polymerization driven thermodynamically by the release of ring strain in a cyclic
monomer. This reaction, under the appropriate conditions with certain catalysts, can
fulfill the requirement of a "living" polymerization, producing polymer chains of a well-
defined molecular weight.59 Acyclic diene metathesis (ADMET) polymerization60'61'62 is
a step-growth equilibrium polycondensation that is driven by the release of a small
molecule (typically ethylene). Essentially, this is self-metathesis with a diene. During
the course of the polymerization, the ethylene condensate is removed, and the equilibrium
is shifted toward the polymer. Acyclic diene metathesis can be used to control precise
functional group or branch placement along a polymer backbone. This has been utilized
in modeling studies of industrial polymers formed by free-radical polymerization
techniques, such as ethylene/propylene copolymers.63'64'65
1.4.1 Historical Overview of the Mechanism
The history of olefin metathesis is quite fascinating.44'45 No one before the 1950s or
early 1960s would have ever predicted that a reaction in which a double bond was
apparently cleaved and its pieces put back together again would be even remotely
possible. The first metathesis publication was in 1960, where the article was on the ring-
opening polymerization of norbornene, using a catalyst derived from titanium
tetrachloride and lithium aluminum tetraalkyls.66 Interestingly, the reaction was termed a
"coordination polymerization" in this article. In 1964, the first journal article on a
metathesis exchange reaction was published, which reported olefinn disproportionation"
of propylene to 2-butene and ethylene.67 Early on, these two fields were researched
independently from one another; the ring-opening polymerizations were typically
initiated with Ziegler-Natta type catalysts, such as MoC15/Et3Al, at or below room
temperature, while the exchange reactions were typically conducted with oxide catalysts,
such as MoO3/A1203, at elevated temperatures.44 It was not until 1967 that Calderon and
coworkers46 recognized that these two reactions were related, and used the same catalyst
(WCl6/EtAlCl2/EtOH in a 1/4/1 ratio) in both reaction types,68'69 beginning the realization
of the diverse opportunity of the olefin metathesis reaction.
The determination of the olefin metathesis mechanism was a controversial problem
of the late 1960s and the 1970s, whereby several different mechanisms were proposed.
The first major mechanism to be proposed was a quasicyclobutane-metal complex
mechanism, shown in Figure 1-14. Although this idea was conceived by Bradshaw and
coworkers,70 the principal publication in support of this mechanism was by Calderon and
coworkers in 1968.71 This mechanism was based on the rules of Woodward and
Hoffman,72 the bisolefin-metal structure proposed by Mango and Schachtschneider
(shown as starting material in Figure 1-14),73 and the observation of a random
distribution of products. The three reported steps in the processes were formation of the
bisolefin-metal complex, transalkylidenation (Figure 1-14), and a rapid olefin exchange
on the metal.71'74
MLn MLn MLn MLn
H2. 2 -H2H \CHR
H2 R H CHR H2 HRH2 RHC HR
1CHR H2C CHR H2C--CHR
Sometimes H2C---CHR H2C:.-;CHR
illustrated as: M or IM
H2C---CHR H2C-- -CHR
Figure 1-14. Early-proposed quasicyclobutane mechanism for olefin metathesis, which
was later found to be incorrect
Although later found to be incorrect, this mechanism was proposed because it did
explained several traits inherent to the metathesis mechanism. For one, it did explain that
the reaction proceeds via scission of the olefinic bond (transalkidenation), as opposed to a
transalkylation, whereby the alkyl substituents on the olefinic carbon atoms exchange
positions instead. A deuterium study was conducted where 2-butene and 2-butene-d8
were mixed together in the presence of catalyst, and the only new product observed was
2-butene-d4.71 If transalkylation were the primary mechanism, -d2, -d3, -ds, and -d6 species
would also have been expected. This mechanism also explains that a random distribution
of products is observed. For example, if 2-pentene is mixed with catalyst, the product is
1:2:1 ratio of 2-butene, 2-pentene, and 3-hexene, respectively. Similarly, this mechanism
also explains the mixture of cis and trans geometric isomers that is observed in products
from pure cis/trans starting materials.
Calderon came to these conclusions since he allowed all his metathesis reactions
reach equilibrium, which is why he observed a random distribution and the mixture of
cis/trans isomers. This mechanism does not account for the appearance of cross products
early on in the reaction. In addition, the quasicyclobutane-metal complex is unconven-
tional bonding theory with no precedent to base this type of reaction intermediate on.
In 1972, a second "pair-wise" mechanism was proposed by Grubbs and coworkers
with a tetramethylene metallocycle as an intermediate, shown in Figure 1-15.7 This
mechanism required the same bisolefin-metal complex as a starting point as in the first
mechanism. It is interesting that Grubbs, a pioneer in the field of olefin metathesis,
proposed this incorrect mechanism. His later work was a huge contribution in providing
evidence of the proper mechanism and still today continues to be a world leader in
metathesis. Grubbs had found evidence of a carbon-metal o-bond, and with similar
o-bonded organometallic intermediates proposed in the work of Katz76 and Eaton,77 this
mechanism seemed reasonable. In his work, Grubbs and coworkers claimed that the
tetramethylene metallocycle was formed by reacting 1,4-dilithiobutante with WCl6,
which would then decompose to form metathesis products.7 However, no proof that the
RHC CH2 R
SM R
RHCO CH2 R R
R R M=CH2
CHR CH2 R M
||I I M I M
CHR CH2
Figure 1-15. Tetramethylene metallocycle mechanism for olefin metathesis, which was
later found to be incorrect
five-membered ring intermediate was actually formed was presented, and the rearrange-
ments observed could have occurred after the olefins were formed.
Later, Grubbs and coworkers argued that these two "pair-wise" mechanisms
(Figures 1-14 and 1-15) would actually not give a random distribution of products.78'79
The case used to demonstrate this was the determination of deuterium content in the
ethylene released from mixtures of 1,7-octadiene and 1,7-octadiene-l,l,8,8-d4. Therein,
they claimed that a "pair-wise" mechanism would have less ethylene-d2 than in a
statistical distribution, which is what is actually observed.78'79
The proper mechanism (Figure 1-16) was first proposed by Herisson and Chauvin8o
in 1970 in a publication that was overlooked by most until the mid-1970s.78'79'81 This
mechanism is distinctly different from the other two because it involves a metal-carbene
complex as the active catalyst, instead of a "pair-wise" interchange. Essential to this
mechanism is the presence of a metal carbene (labeled as the initiator or methylene
alkylidene in Figure 1-16). Retrospectively, in the days before well-defined catalyses,
the carbene must have formed in situ through unknown mechanisms. The metal carbene
reacts with olefins to form a metallocyclobutane complex. This complex is similar to the
one shown in Figure 1-15 except that it is a four-membered ring instead of five. The
metallocyclobutane can then break apart, reforming a metal carbene and an olefin. This
carbene can then continue in the catalyst cycle and reacts with other olefins in the
solution.
This process of the forming and breaking apart of a metallocyclobutane complex is
illustrated in each line of the mechanism in Figure 1-16. In line (a), the initial alkylidene
CHR' CHR R'HTCCHR R'HC=CHR
(aII + II +
M CH2 M-CH2 M=CH2
initiator starting methylene
olefin alkylidene
M=CH2 M-CH2 M CH2
(b) + II + II
RHC=CH2 RHC CH2 CHR CH2
more starting new
olefin product 1
M CH2 M-CH2 M=CH2
(c) II + II +
CHR CHR R'HC-CHR RHC=CHR
more starting new
olefin product 2
Statistical Mixture of:
Overall M=CHR
Reac R'HC=CH2 R H2C=CH2 + R'HC=CH2 + R'HC=CHR'
Reaction:
new starting new
product 1 olefin product 2
Figure 1-16. Metal-carbene mechanism for olefin metathesis, which is now well
established as the correct mechanism
(labeled as the initiator) reacts with an olefin to form a methylene alkylidene. This
carbene can then continue along in the catalyst cycle as the catalytically active species.
Eventually new products are formed from the pieces of the original olefin, as shown in
lines (b) and (c) of Figure 1-16. A random statistical distribution will result at
equilibrium since all of these reactions are reversible. There is a small amount of a by-
product (R'HC=CHR) that is formed from the initial alkylidene in line (a). However,
since normally only a catalytic amount of the initiator is added, the amount of this by-
product is usually negligible. In the modern times of well-defined catalysts, this
mechanism is represented more appropriately as a catalytic cycle with an initiation
period, as is shown for an example of RCM in Figure 1-17. This more clearly illustrates
that the methylene carbene is the active species that is continually being regenerated as
the reaction progresses.
starting
olefin
MLn
olefin R
initiation Metathesis
LnM= LnM=CH2 Catalytic
R Cycle ML
product
MLn
Figure 1-17. Catalytic cycle for olefin metathesis in an example of RCM
The metal-carbene mechanism was the only one of the three proposed mechanisms
that did not require unusual bonding theory in an intermediate of the mechanism. Also
unlike the other two, this mechanism accounted for the appearance of cross products in
the early stages of the reaction. Katz and coworkers reexamined the reaction of
unsymmetrical acyclic olefins with cyclic olefins and found the cross-coupled products
appeared before a significant buildup of the scrambled acyclic olefin, which supports the
metal-carbene mechanism.81 Grubbs' deuterium work in refuting a "pair-wise" mech-
anism also supported this mechanism.78'79 Nearly 30 years later, the metal-carbene
mechanism has become well established as the proper mechanism for olefin metathesis,
where well-defined metal-carbene catalysts have found wide use in the diverse field of
olefin metathesis.
1.4.2 Development of Well-Defined Catalysts
The most significant thing to occur in the field of olefin metathesis in the past 20
years is the development of well-defined catalysts. These well-defined carbene structures
tend to act more as initiators rather than catalysts.44 They offer a better understanding of
the mechanism, a much higher degree of control, and an allowance to tailor the catalysis
to meet the needs of the application. Each catalyst offers its own advantages and
disadvantages, where a varying degree of metathesis activity, catalyst lifetime, functional
group tolerance, and unwanted side reactions are possible. A few particular catalysts, due
to their combination of traits and availability, have become tremendously popular, and
have found wide use in the field.
The first well-defined alkylidene catalysts that were active towards olefin
metathesis were developed by Schrock and coworkers; in 1986, tungsten-based
complexes were synthesized,82 followed by molybdenum-based complexes the following
year.83 These classes of catalysts have recently been reviewed in the literature with a
comprehensive list of all the tungsten and molybdenum alkylidene complexes that have
been synthesized.84 Common versions of these catalysts are shown as complex 1 and 2 in
Figure 1-18. These high-oxidation state, 14-electron complexes are pseudotetrahedral,
electron-deficient, and only contain sterically crowded, covalently bound ligands
including the alkylidene. This steric crowding accounts for the catalyst's stability toward
bimolecular decomposition, which was a problem in the development of these catalysts.84
Schrock's catalysts are extremely active. However, strict anhydrous and inert atmos-
pheric techniques must be employed for their use, and they are not very tolerant of
functionality on the olefin substrate.
In any catalyst system, functional groups on the substrate or solvent (including
oxygen and water) can interfere with the catalytic process.85 If other functionality in the
system can bind competitively to the active metal center, the catalyst can be deactivated.
Another possibility is that the functionality can react directly with the metal center,
causing decomposition of the catalyst.
N N
I' mP h IF \ 0 / II p P
F3c- w Ph F3C 'Mo Ph
F3C F3C
0 0
F3C CF3 F3C CF3
1 2
Figure 1-18. Common version of Schrock's tungsten (1) and molybdenum (2) based
olefin metathesis catalysts
There were several early reports in the 1960s that the ruthenium complex,
RuC13(hydrate), could perform ROMP in refluxing ethanol and aqueous emulsion
conditions.86'87 The yields for polymer were low, but the fact that metathesis was
occurring in a protic environment at all is quite significant. Table 1-1 shows the binding
trends that titanium, tungsten, molybdenum, and ruthenium tend to follow. Unlike the
other three, ruthenium reacts preferentially with olefins over most other species. This
gives the ruthenium catalysts stability towards alcohols, aldehydes, carboxylic acids, and
water that the Schrock's catalysts do not possess. Regrettably, ruthenium was not
seriously considered as a candidate for well-defined catalysts for more than two decades,
most likely because of the discouragingly low metathesis activity of ruthenium salts.
In the late 1980s, the potential of ruthenium catalysts were reexamined. It was
found that indeed RuC13(hydrate) did catalyze ROMP; however long initiation times were
observed in organic solvents.88'89 When strict anhydrous conditions were used, the
initiation rates were even slower. Unexpectedly, it was found that ROMP initiated rather
quickly in aqueous conditions.90 Water was found, not only to be compatible with the
ruthenium tetrachloride catalyst system, but also it was actually beneficial to the initiation
process.
Table 1-1. Functional group tolerance of early and late transition metal olefin metathesis
catalysts
Titanium Tungsten Molybdenum Ruthenium
Acids Acids Acids Olefins
Alcohols, Water Alcohols, Water Alcohols, Water Acids
Aldehydes Aldehydes Aldehydes Alcohols, Water Increasing
Reactivity
Ketones Ketones Olefins Aldehydes Reactivity
Esters, Amides Olefins Ketones Ketones
Olefins Esters, Amides Esters, Amides Esters, Amides
Since this period, a number of well-defined ruthenium catalysts have been
developed.91 A select group of theses are illustrated in Figure 1-19. The breakthrough
came in 1992, when Grubbs and coworkers developed catalyst 3, the first ever metathesis
active ruthenium alkylidene.92 They applied a methodology to ruthenium that had been
successful in the synthesis of tungsten alkylidenes, whereby 3,3-disubstituted cyclo-
propenes were used as carbene precursors.93 With this catalyst, norbonene could be
polymerized in the presence or absence of water or ethanol. For the first time, a
propagating alkylidene could actually be observed by 1H NMR during the course of a
metathesis polymerization.
Although these observations were remarkable, catalyst 3 was limited to ROMP of
highly strained cyclic monomers. A trend observed with early transition metal catalysts
was that metathesis activity increases with electron-withdrawing ligands; thus the
triphenylphosphine (PPh3) ligands were exchanged with other phosphine ligands of lower
basicity in an attempt to improve the catalyst system. After months of testing these new
catalysts without success, it was found that one of the largest and most electron-donating
phosphines available had a tremendous increase in metathesis activity.94 This ligand was
tricyclohexylphosphine (PCy3), and by replacing the PPh3 ligands to yield catalyst 4, it
was learned that the larger and more basic the phosphine ligands on ruthenium, the higher
PR3 PCy3 N
CI// Cci
'Ru Ph C'Ru ClRu P
CIf Ci^t Ph ci Ph
PR3 Ph PCy3 PCy3
R=Ph 3
5 6
R = Cy 4
SrI 1 ") "i
CIu R CI/
CI1 CIitf Ph CI
0 PCy3 0
7 8 9
Figure 1-19. Select well-defined ruthenium metathesis catalyst
the metathesis activity, contrary to the thinking of the ligands on the Schrock's catalysts.
The a-donation of the phosphine ligand to the metal center is thought to promote the
formation of the mono-phosphine complex through the dissociation of the other
phosphine ligand and stabilization of a vacant site in the trans-position. This is known to
be an important step in the mechanism with ruthenium alkylidene catalyst.95'96'97 When
this was initially examined, the dissociation of a phosphine ligand was surprising since it
meant that the mechanism contains electron-deficient, 14-electron, metallocyclobutane
intermediates.95 In addition, the o-donation of the remaining phosphine to the metal helps
stabilize these electron-deficient intermediates. Unlike its predecessors, catalyst 4 could
perform ROMP of low ring-strain monomers, such as cyclopentene, and it became the
first ruthenium catalyst that was active towards acyclic metathesis processes. Due to its
robustness in air as a solid and its functional group tolerance, this catalyst could be used
in a wide range of synthetic applications. However, it did require the uses of
diphenylcyclopropene in synthesis, which limited its availability.
In 1995, a new methodology for synthesizing substituted ruthenium alkylidenes
was developed using diazoalkanes as the carbene source.98'99 This methodology led to the
synthesis of several more ruthenium alkylidenes, the most active of which was the
benzylidene catalyst 5. The activity of this type of catalyst is highly dependent on the
identity of both the X- and L-type ligands, as well as the alkylidene. Whereas the activity
increases with larger, more electron-donating phosphines (PPh3 << PiPr3 < PCy3), the
activity decreases with larger, more electron-donating halogens (Cl > Br >> I).95 Since
incoming olefins may bind trans to a halide, the more electron-donating halides will
weaken the ruthenium-olefin bond and disfavor olefin coordination, thereby inhibiting
the activity of the catalyst. The more electronegative chloride ligands are the most active
halogen choice. Also, the alkylidene plays a role in the initiation of the catalyst. Alkyl-
substituted alkylidenes display more efficient initiation rates than the methylidene version
of the catalyst, presumably by assisting in phosphine dissociation. Even faster initiating
are ester-substituents; however, these catalysts are not very stable. The benzylidene turns
out to be an intermediate case where the phenyl ring is somewhat electron donating, but it
is large enough that it can assist in phosphine dissociation. The benzylidene catalyst (5)
can undergo metathesis with ethylene at room temperature to quantitatively form the
methylidene complex within minutes.85
Known today as the "first-generation Grubbs" catalyst (5), this 16-electron
ruthenium alkylidene is the combination of the bulky, electron-donating PCy3 ligand, the
more electron-withdrawing halide ligands, and the fast-initiating benzylidene moiety. It
exists in a distorted square pyramidal geometry with the alkylidene in the axial position
and alternating phosphines and chlorines in the equatorial positions.99 With its high
activity and additional syntheses that have been developed from readily available starting
materials,100,101,102 the first-generation Grubbs catalyst has found extensive use in the vast
field of olefin metathesis.
In 1999, a new generation of ruthenium-based olefin metathesis catalysts bearing
N-heterocyclic carbene (NHC) ligands in place of one of the phosphine ligands was
developed, again by Grubbs and coworkers.103 The most notable of these is catalyst 6,
which has become known as the "second-generation Grubbs" catalyst. N-Heterocyclic
carbene ligands are stronger o-donors and much less labile than phosphines; therefore
these ligands are not able to readily dissociate from the ruthenium as the phosphine
ligands do. Earlier catalysts with two NHC ligands were found to have little
improvement over the activity of catalyst 5.85 However, a mixed-ligand system, as in
catalyst 6, was designed in order to utilize the stronger electron-donating ability of the
NHC ligand to enhance the dissociation of the more labile phosphine ligand in the trans-
position. In addition, the steric bulk of the NHC ligand, as well as it electron-donating
properties more effectively stabilize the electron-deficient, 14-electron intermediates that
promote olefin metathesis. After exploring different substitution patterns on the NHC
ligand, it was found that mesityl substitution worked well.103'104 It is thought that the
mesityl substituents provide the metal with considerable steric protection, which
contributes to its high thermal stability.
Before this class of catalysts emerged, it appeared that the gain of functional group
tolerance came at the expense of activity. However, catalyst 6 rivals the activity of
Schrock's molybdenum catalyst (2), but maintains the functional group tolerance of 5.
Catalyst 6 has superior activity to its predecessor (5), which includes high reaction rates
for ROMP of low-strain cyclic olefins, 105 ROMP of the sterically hindered trisubstituted
olefins,105 and RCM of sterically demanding dienes to make tri- and tetrasubstituted
olefins.103'104 Also, for the first time, trisubstituted olefins could be yielded from CM,106
and olefins with a deactivating group directly on the double bond could undergo RCM
and CM.107 Catalyst 6 has high thermal stability at elevated temperatures, and has been
found to have a longer catalyst lifetime than the parent complex 5. Whereas the catalyst
half-life of the methylidene form of the first-generation catalyst is 40 min at 55 C in
benzene,108 the second-generation catalyst has a half-life of 6 h under the same
conditions.103 On the other hand, contrary to what was expected, catalyst 6 does initiates
slower, most likely due to slower phosphine dissociation.85
These 16-electron ruthenium catalysts could conceivably undergo initiation via two
general pathways, an associative route or a dissociative route.96'97 Figure 1-20 shows both
these general pathways. In an associative pathway, the olefin coordinates to the metal
first at a vacant site before the phosphine ligand leaves, thereby making an 18-electron
intermediate (A). On the other hand, in a dissociative pathway, the phosphine ligand
leaves first, leaving a 14-electon complex (B) and another vacant site where the olefin
can then coordinate. Regardless of the order in which this occurs, once the phosphine has
left, and the olefin has coordinated, the resulting 16-electron complex (C) can form the
metalocyclobutane complex and proceed via the mechanism in Figure 1-16 and 1-17.
Recent experimental data96 and density functional calculations97 suggests that it a
dissociative pathway where the phosphine departure is the first step of the initiation.
Associative Pathway:
L L L
l R' + olefin R' PCy3 R' p
CRuI CI2Ru C12Ru L = PCy3
SCI olefin + PCy3 (5)
PCy3 PCy3 R R
(A) (C) or
Dissociative Pathway:
L L L
CI, u R' PCy3 R' + olefin R' N
'Ru-/ Ru- CI2Ru (-
t CI + PCy3 olefin (6)
PCy3 CI R
(B) (C)
Figure 1-20. Associative and dissociative pathways of first- (5) and second- (6)
generation Grubbs catalyst
Other notable "second-generation" ruthenium alkylidene catalysts that have been
employed in this body of work are Hoveyda's catalyst (7),109 Mol's catalyst (8),110,111 and
the Hoveyda-Mol hybrid (9).112 On Hoveyda's catalyst, the alkylidene has an isopropyl
group in the ortho position of the aromatic ring, on which the oxygen atom coordinates to
the ruthenium center in place of the phosphine ligand. Figure 1-21 illustrates how
isopropoxystyrene is released and then returned to the ruthenium complex. This in turn
allows for the recovery of the catalyst with its original alkylidene moiety attached to the
ruthenium center. The oxygen coordination is quite labile while the alkylidene is
attached, which gives it an extremely long catalyst lifetime in solution. However, in
metathesis processes where the isopropoxystyrene dissociates, the catalyst can
decompose if there is nothing to stabilize the 14-electron intermediates.112 This is likely
due to molecular isopropoxystyrene not being very labile when not bound to the metal as
an alkylidene and therefore cannot recoordinate with the ease it does when it is attached.
Fortunately, most metathesis processes where this would be used have some functionality
on the substrate or on the solvent (THF has been reported112 to work well) that can
L = PCy3 or %N N
S(7)
L L
Ru L L c(7
CIO CI CI 2 'Ru -
0 Ru Ru-
Release Return
Figure 1-21. Hoveyda's catalyst undergoing RCM, whereby the original oxygen-
coordinated catalyst is reformed
coordinate and fulfill this purpose. Mol's catalyst further increases the steric bulk of the
NHC ligand with bulky isopropyl groups on its aromatic rings, which allows for
enhanced rates of reaction, particularly at lower temperatures. Apparently, the interaction
of the isopropyl groups with the alkylidene and/or the chlorides promotes phosphine
dissociation. Although this bulkier NHC ligand was originally thought to possibly
increase catalyst stability,"1 it was found to have the opposite effect. At monomer to
catalyst loadings of 450:1, Mol's catalyst has been reported to not go beyond
trimerization under ADMET polymerization conditions at temperatures of 45 OC and
higher due to catalyst decomposition.112 Other catalysts have been developed with a
varying degree of bulkiness on the NHC ligand. In general, the bulkier the NHC ligand,
the more active the catalyst is, but the faster the catalyst decomposes. The Hoveyda-Mol
hybrid, developed within the Wagener research group, combines the thinking from the
two variations on Grubbs' second-generation catalyst, gaining the advantages of both in
one catalyst.
These second-generation catalysts do have a major disadvantage that their
processors do not. This is that olefin isomerization has been observed in many
metathesis processes.113,114,115,116 Although there are examples of isomerization using
earlier metathesis catalysts,113'117 only trace amounts were observed and only after long
reaction periods. The extent to that which isomerization occurs with catalyst 6 and other
ruthenium catalysts with NHC ligands is much higher. For cases where the metathesis
reaction is unfavorable, then isomerization may even be the primary mode of reaction
occurring, as was the case in one study,115 where as much as 100 % yield of the
isomerization by-product was reported in an attempted RCM reaction to make an eight-
membered ring. In an isomerization study from within our group that simulated ADMET
conditions, Lehman and coworkers found significant isomerization of linear hydrocarbon
species occurs at temperatures 50-60 C.113 For one example therein, catalyst 6 was
added to the 14-carbon species, 7-tetradecene, at 55 C. If no isomerization were to
occur, the expected product would be starting material since all possible metathesis
routes are unproductive. Instead, a bell curve of species ranging from nine to twenty
carbons was observed by gas chromatography, centered around the 14-carbon species.
These other products could only appear if isomerization were occurring. Also reported
therein is that the extent of this isomerization with catalyst 6 is greatly reduced, but not
completely suppressed, at temperatures 30 C and below.
Alkene isomerization can occur by two known mechanisms, the metal-hydride
addition-elimination mechanism and the 7r-allyl metal-hydride mechanism.118'119 It is not
completely clear if the isomerization in metathesis reactions is promoted from the catalyst
itself, decomposition products, or impurities in the catalyst synthesis, although even from
early on, it has been thought that the most the likely culprit is a ruthenium hydride
complex that forms from the decomposition of the catalyst, which would presumably
isomerize through the metal-hydride addition-elimination mechanism. Ffirstner and
coworkers reported the isolation of a ruthenium dihydride complex RuCl2(PCy3)2(H)2
during the synthesis of catalyst 8, which could be a cause.110 The key to understanding
isomerization may lie in understanding the pathway of catalyst decomposition. Although
this pathway is not totally clear to date, some things are known about the relative rates of
decomposition. Firstly, the rate of decomposition is second order in respect to catalyst
concentration, whereby two ruthenium carbenes likely form a bimetallic species as a by-
product. Also, the rate is retarded with the adding of extra phosphine ligand, which
implies that the phosphine-dissociated intermediate is the species that likely couples
together. This is consistent with the fact that catalyst lifetimes are shorter with catalysts
that have fast phosphine dissociates rates. Most recently, Grubbs and coworkers may
have solved this mystery by isolating structure 10 (Figure 1-22) directly from catalyst 6
under metathesis reaction conditions.120 An in depth isomerization study with this
diruthenium hydride by-product has not yet been conducted; however, they did report one
S55 OC M es c
2 RI -55-0 N Ru-C Ri--CI + 2 CH3PCy3+CI
Cf Ph C6H6 H r
PCy3 N H
Mes
6 10
Figure 1-22. Bimolecular decomposition of catalyst 6 to form diruthenium complex 10,
which is likely responsible for isomerization during olefin metathesis
example of isomerization of an olefinic species using this species (10). At this time, this
bimetallic hydride complex appears to be the most likely culprit responsible for the
isomerization that occurs during olefin metathesis of second-generation Grubbs catalyst
with similar complexes possible for other second-generation catalyst. Also proposed
therein is a mechanism of decomposition to arrive at this structure (Figure 1-23). 120 This
mechanism is not only via a bimolecular pathway, but it is also consistent with phosphine
dissociation being an important step in the decomposition of the catalyst. If this new
understanding of catalyst decomposition is correct, it may now be possible to design a
catalyst with the inhibition of this mechanism in mind.
Even though olefin metathesis is a latecomer to the modern chemical scene, it has
flourished into a widely used reaction in organic and polymer chemistry. Each new
catalyst that is developed is another tool available to the metathesis world, each with its
own advantages and disadvantages. The more that is understood about the mechanisms
that these catalysts undergo, the higher potential there will be for synthesizing new
catalysts that are tailored to exploit desirable properties.
CRu =CH2 CI2Ru=CH2 + PCy3
Ct
PCy3
6
N%- N N -%N Cy3P CH2
S CH2
CRu-CH2 Ruj Cl Ru-
Cl, PC3 P3 CI' PC
N I Mes CN/ N
C\'Ru + Cl2Ru=CH2 -N Ru Ru CH2
Cl (' cl 'cIc l
CICN H
Mes
y M es N M es N \ i
N R Ru CH N Ru-CRu CI
N H CCl N H 0
Mes Mes
10
Figure 1-23. Mechanism proposed by Grubbs for the decomposition of catalyst 6
1.5 Opportunities of Olefin Metathesis in the Solid State
The mission of this work is to demonstrate that a particular mode of olefin
metathesis, ADMET, can operate in the solid state within the same principles as
semicrystalline condensation polymers. Acyclic diene metathesis, RCM and CM all
involve condensation steps, so in principle all of them could operate in the solid state,
assuming some degree of amorphous character is present during the conversion. Of the
three, ADMET chemistry leads to macromolecules, and like the industrial polymers that
are typically performed as a solid-state polymerization polyesterss, polyamides, and
polycarbonates), the equilibrium is shifted towards the polymer by the removal of small
condensate. In most cases of ADMET, the condensate that is released is ethylene, as
shown in the ADMET mechanism60'121 in Figure 1-24.
LnRu
R R ethylene
is
olefn released
olefin
initiation R' ADMET
LnRu\ -R LnRu CH2 Mechanism R RuLn
R R
R
n LnRu
polymer n R R
Figure 1-24. The ADMET mechanism whereby the equilibrium is shifted toward the
polymer by the release of ethylene
Solid-state ADMET has a distinct advantage over common solid-state polyester,
polyamide, and polycarbonate condensation reactions, and that lies in the volatility of the
condensate. Table 1-2 compares the condensate boiling points of typical industrial SSP
and ADMET. All other known SSP examples release high boiling condensates such as
ethylene glycol (PET), water (Nylon), or phenol (polycarbonate), all requiring reaction
temperatures of greater than 100 C to drive off the condensate. In contrast, the
condensate for ADMET boils at -104 OC, and therefore can be conducted at significantly
lower temperatures. Fortuitously, this allows for utilization of reaction temperature
where metathesis catalysts are typically run (30-60 C). In addition, lower temperatures
are more economically viable and cause fewer side reactions to occur. This process
could possibly lead to the generation of highly pure, high molecular weight polymers at
or near room temperature. Further, the process may offer a new route to the direct
preparation of intractable polymers, which presently are made via indirect techniques.
Table 1-2. Comparison of condensate boiling points
Condensate
Polymer Type Example Condensate n
Boiling Point
ADMET Polyoctenylene Ethylene -104 C
Polyester PET Ethylene glycol 198 C
Polyamide Nylon 6,6 Water 100 C
Polycarbonate Lexan Phenol 182 C
Since this work is conducted in the solid state, Chapter 2 is a brief explanation of
the experimental methods, uncharacteristic of typical organic synthesis, which were
developed and used throughout the work. Chapter 3, which contains most of the work
performed for this dissertation, focuses on ADMET polymerization of semicrystalline
polymers in the solid state in an effort to mimic the solid-state polycondensation
processes conducted in industry. Chapters 4 and 5 are extensions of this work aimed at
more unconventional principles, whereby Chapter 4 concentrates on solid-state ADMIET
of rigid-rod polymers and Chapter 5 focuses on attempting ring-closing metathesis in the
solid state.
CHAPTER 2
EXPERIMENTAL METHODS
Since it is the objective of this work to conduct reactions in the solid state,
unconventional organic methods had to be applied for the solid-state reactions. It is
therefore appropriate to devote a brief chapter on the techniques that were used and
developed throughout this work. This chapter will describe the typical solid-state
apparatus, the different methods used to make a solid mixture of diene and catalyst, the
process of sampling the reaction, and then the characterization techniques that were used
on the samples throughout this work.
2.1 The Solid-State Apparatus
The typical experimental setups were trivial. The apparatus consisted of equipping
a 50 mL three-neck round bottom flask with two vacuum adapters (one requiring a
stopcock) and a stopper (all 24/40 joints). The glassware was typically dried overnight in
the oven at 80 C. While still hot, the apparatus was assembled whereby the vacuum
adapter without a stopcock is connected to a high vacuum/argon line and the vacuum
adapter with a stopcock is connected to a bubbler as shown in Figure 2-1. With the
stopcock closed, the apparatus is evacuated and then backfilled with argon three times.
The line to the bubbler was also purged with argon during the backfilling process. The
apparatus was then flame dried and allowed to cool under vacuum. At this point,
reactants could be added with positive argon pressure. Reaction temperature was
maintained in an oil bath (+/- 2 C) by one of three temperature controllers. The first was
a J-KEM Scientific Model 210 with a K-probe (-200 to 1372 C) and a heating coil. The
second was a Barnant Company Model 689-0000 with a T-probe (-200 to 400 OC) and a
heating coil. The third was a Corning Temperature probe #400188 for Coming PC-420
hot plates.
vacuum bubbler
argon
solid
temperature-controlled oil bath
Figure 2-1. Drawing of a typical apparatus for solid-state reactions (For experiments
with argon flow, the stopcock was left open to allow air flow to the bubbler.
Likewise, the stopcock was closed for reactions subjected to vacuum.)
The reactor was then left intact at the set temperature with argon flowing over the
solid solution to the bubbler throughout the course of the reaction. In some cases vacuum
was used with the stopcock closed instead of argon flow. Typically these experiments
were run for 7 d and sampled periodically at specified time intervals.
2.2 Making a Solid Mixture of Reactants
Multiple methods of creating a "solid solution" of reactants were experimented
with. This section aims to convey these methods in detail. The most common method
used in this work was an initial polymerization in the melt state before subsequent
crystallization to a solid mixture of prepolymer and catalyst. For polymerizations aimed
to start from monomer, as well as reactions aimed to conduct RCM in the solid state,
other means of creating a solid mixture were sought out. The freeze-drying of the solid
monomer and catalyst together was utilized for these purposes using benzene as a
solvent. In later studies of the same type, it was found that simply grinding the solid
diene and catalyst together in a mortar and pestle could produce reaction mixtures that
achieved similar molecular weight increases. This grinding technique was also applied to
reactions mixtures of cleaned up (precipitated to remove catalyst) prepolymers and
catalyst for additional polymerization in the solid state.
In making up these solid solutions, scales of 0.2 g to 0.5 g of monomer were
typically employed with catalyst loadings typically from 100:1 to 500:1 monomer to
catalyst. The details of what was used for each particular experiment including catalyst
loading and reaction temperature are specified in the results and discussion section as
each experiment is conveyed and appears in the results and discussion section of the
individual chapters where they are presented.
2.2.1 Melt Polymerization
This method most closely simulates industrial SSP by beginning the polymerization
in the melt state with subsequent crystallization of the growing macromolecule and
continued polymerization in the solid state; therefore this method was most often used in
this work. The melt portion of this procedure was conducted either at the same
temperature as the solid-state portion, or at elevated temperatures to keep the forming
polymer in the melt for a period of time. In the former case, the reaction mixture
solidified after an amount of time at a constant temperature, after which the solid mixture
was first sampled to mark the start of the solid-state reaction, which was then continued
at the same temperature. In the latter case, elevated temperatures were needed to keep
the monomer and subsequent polymer in the melt for an amount of time before the
eventual crystallization by cooling to a more appropriate reaction temperature. In further
attempt to simulate the industrial process, the solid was often taken out of the reaction
flask, and broken up into small particles in order to increase surface area to assist in
condensate release. This was usually accomplished by grinding with a mortar and pestle;
however, this was sometimes replaced by or complemented with the breaking up of the
solid with a spatula. The small particles were then transferred to a fresh apparatus at the
desired temperature for continued reaction in the solid state.
2.2.2 Freeze-Drying
For polymerizations aimed to start from monomer and reactions aimed to conduct
RCM in the solid state, freeze-drying of the two reactants together was utilized. Separate
solutions of the monomer (or RCM diene) and catalyst were prepared in minimal
amounts of benzene and cooled in an ice bath. Usually one or both of these solutions
would actually solidify depending on concentration. A similar reaction flask to that
shown in Figure 2-1 was setup except that both vacuum adapters required a stopcock.
One vacuum adapter is connected to a trap, and the whole system is evacuated and flame
dried as described above. The two cooled benzene solutions were then combined in the
reaction flask with positive argon pressure, and immediately frozen (freezing point of
benzene = -5.5 C) in a dry ice/acetone bath (-78 C). Once frozen, the stopcock to the
trap was closed, and quickly the trap was evacuated and submerged in a dewer containing
liquid nitrogen. At this point the reaction flask was also subjected to vacuum by opening
the stopcock on the vacuum adapter, subliming the benzene from the reaction flask and
into the trap with liquid nitrogen as shown in the setup in Figure 2-2. A solid mixture of
diene and catalyst results after all the benzene has been removed. The amount of benzene
collected in the trap was typically very close to the combined amount from the two
original solutions (within 1/2 mL for 6 mL of benzene). The complete removal of
benzene and the possible presence of any conversion from the short time in the liquid
state were always checked by NMR spectroscopy.
Vacuum
dry ice/acetone bath
liquid nitrogen
Figure 2-2. Freeze-drying apparatus with liquid nitrogen trap
2.2.3 Grinding Together Diene and Catalyst
A solid mixture of diene and catalyst could also be made simply by grinding the
solid diene and catalyst together with a mortar and pestle. In several cases, the starting
materials would stick to the mortar and the pestle; therefore the material was scraped off
the sides of the mortar with a spatula after each grinding step. This process of grinding
and scraping was completed multiple times in order to achieve a more homogeneous
mixture. Three grinding and scraping steps were performed for the making of each solid
mixture to be consistent throughout these studies. For the best possibility of obtaining a
homogeneous mixture, the two solid starting materials were mixed with a spatula prior to
the grinding process to prevent clustering of the catalyst in any one part of the solid
mixture.
2.2.4 Regrinding Clean Polymer with Fresh Catalyst
This grinding technique was also applied to reaction mixtures of clean prepolymers
and catalyst. Here a prepolymer is first synthesized under melt or solution conditions that
will be specified in the individual chapters. The resulting prepolymer was then quenched
(see next section) and precipitated (usually in cold methanol) to remove the catalyst.
After filtration and drying under high vacuum, the prepolymer could then be ground
together with fresh catalyst as described in the previous section.
2.3 Sampling of the Solid Mixtures
Since solid-state reactions are inherently slow, each experiment was run for an
extended period of time (usually 7 d). The reaction progress was monitored by opening
up the reaction flask with positive argon pressure, and scoping out a small portion of the
solid mixture with a spatula. This solid could then be used for molecular weight
characterization in order to observe the progress of the reaction.
In order to prevent further reaction in the solution states of the characterization
solvent, each sample had to be quenched. This was accomplished by the addition of ethyl
vinyl ether (EVE), a known inhibitor of olefin metathesis with the Grubb ruthenium
catalysts.94,122,123 These ruthenium-carbene complexes reaction rapidly and quantitatively
with EVE to form a Fisher carbene as shown in Figure 2-3. The resultant complexes
have been shown to be metathesis inactive at standard conditions and only weakly active
at elevated temperatures. The addition of EVE was accomplished in one of two ways.
Either it was added via a pre-made quench solution, or added directly to the sample.
LnRu=CH + O -- LnRuo--- + H2C=CH
R R
metathesis inactive
R = H or polymer ethyl vinyl ether metathesis inactive
Fisher carbene
Figure 2-3. Reaction of ruthenium catalysts with EVE
Another concern was the possibility of radical chemistry occurring while in
solution for samples that had to be stored for an extended period of time prior to analysis.
This was addressed by the addition of a free radical scavenger, 2,6-di-tert-butyl-4-
methylphenol, commonly know as butylated hydroxy toluene (BHT). The mechanism by
which BHT inhibits radical chemistry is shown in Figure 2-4. A hydrogen radical is
transferred from the phenolic oxygen to quench the radical species. The resultant free
radical form is stable because the radical moiety is surrounded by the two bulky tert-butyl
groups and therefore cannot react in typical radical chemistries. In experiments where
nuclear magnetic resonance (NMR) spectroscopy was the primary mode of molecular
weight determination, BHT was not be used because the OH peak of BHT (5.0 ppm in
CDC13) overlapped with one set of the external olefin peaks from typical ADMET dienes
(typical meaning the monomers have long-methylene spacers between the olefins and any
functionality present). In these cases, the need for accurate integration outweighed the
need to prevent radical chemistry. Between storing samples in a refrigerator and taking
the measurements as soon as possible, prevention of radical chemistry was sufficiently
achieved.
S / + R \ / + R-H
Figure 2-4. Free radical scavenging with BHT
For when the 7-d reaction time is completed, quench solutions were prepared with
the aim of quenching a 0.5 g batch of polymer, which would contain approximately 5 mg
of catalyst at 500:1 monomer to catalyst ratio, with 0.5 mL of solution. After adding
EVE and BHT to a volumetric flask, the remaining volume was filled by the GPC
solvent. These quench solutions contained EVE to catalyst ratios of 100:1 to 500:1 and
BHT loadings of 0.1 % to 5.0 % by weight.
When making up individual samples for GPC, the polymer samples (-5 mg) were
first quenched by adding 2-5 drops of the proper quench solution, and then diluted to the
desired concentration with the GPC solvent. Early NMR experiments were quenched in
the same manner until it was discovered that the OH-peak was interfering the external
olefin integration. For the NMR experiments where BHT was excluded, which is the
case for the majority of the NMR data presented in this work, only EVE was used when
quenching the samples. For NMR in chloroform-d (CDC13), batch quench solutions were
prepared simply by adding 1.0 mL of EVE to a 50 g bottle of CDC13. This is an
approximate 100-fold excess of the amount suggested of EVE (100:1 EVE to catalyst) to
quench a 20 mg sample with the 1.0 mL portion needed for the NMR in a 300:1
monomer to catalyst experiment. Storing these solutions in the refrigerator is imperative
because EVE decomposes (one by-product observed was acetylaldehyde) in the slightly
acid chloroform environment at room temperature. For all other NMR solvents, EVE
(2-3 drops) was added directly to the sample with subsequent addition of approximately
1 mL of the NMR solvent. This was found to give approximately the same EVE
concentrations as the CDC13 solutions above.
2.4 Characterization Techniques Used on the Samples
This section details the general modes of polymer characterization that were used
on samples throughout this work. In particular, gel permeation chromatography (GPC)
and/or NMR end-group analysis were used at the two primary forms of molecular weight
determination. Other forms of characterization are detailed within the individual
chapters.
2.4.1 Gel Permeation Chromatography (GPC)
Two different GPC systems were used throughout this work. The first machine
was a modular system with Waters U6K liquid chromatograph equipped with an
Eppendorf TC-45 column heater, Polymer Laboratories Ultrastyragel C linear mixed bed
columns, a Hewlett Packard HP 1047 refractive index detector, and an ABI Analytical
Kratos Division Spectroflow 757 ultraviolet detector. On this machine, relative
measurements were taken relative in polystyrene in HPLC grade tetrahydrofuran (THF)
of cholorform at 1.0 mL/min with sample concentrations of 4.0 to 8.0 mg/mL. The
second machine was a Waters GPCV 2K with high temperature and light-scattering
capabilities. Relative measurements were taken using HPLC grade THF at 40 OC on
Waters Styragel HR 5E columns relative to polystyrene standards at 1.0 mL/min. The
light-scattering equipment was a Wyatt Technologies three-angle miniDAWN that was
inserted into the heated compartment of the GPCV 2K. On this machine, sample
concentrations of approximately 0.5 mg/mL were used. All relative GPC for both
machines were calibrated against ten narrow (PDI > 1.07) polystyrene standards
(Polymer Laboratories) ranging in number average molecular weights ranging from
580 g/mol to 275,200 g/mol.
2.4.2 Nuclear Magnetic Resonance (NMR) Spectroscopy
The spectra were all taken on a Varian Gemini 300 MHz, a VXR 300 MHz, a
Mercury 300 MHz, or Mercury 300 MHz broadband. In CDC13, 1H spectra were
referenced to TMS as 0.0 ppm and 13C spectra were reference to CDC13 (t) as 77.23 ppm.
In Toluene-ds, 1H spectra were referenced to TMS as 0.0 ppm and 13C spectra were
reference to toluene-d8 (s) as 137.86 ppm. In DMSO-d6, 1H spectra were referenced to
DMSO-d6 (5) as 2.50 ppm and 13C spectra were reference to DMSO-d6 (7) as 39.51 ppm.
In D20, 1H spectra were referenced to the HOD peak as 4.80 ppm and 13C spectra were
not referenced. These reference numbers are all from a Cambridge Isotope Laboratoiries,
Inc. NMR solvent data chart.
End-group analysis was performed by comparing the integration from internal
olefin peaks (5.28 to 5.46 ppm) and the external olefin peaks (5.72 to 5.90 ppm and 4.88
to 5.06 ppm). These chemical shifts are of typical polymers made via ADMET that have
long-methylene spacers between the olefins and any functionality that may be present in
the monomer when taken in CDC13 relative to TMS. For an ADMET polymer with a
degree of polymerization (DP) of 1 + n (Figure 2-5), the signal appearing from the
internal olefins is from 2n protons per chain, while the signal appearing from the external
olefins is from six protons per chain. Therefore by comparing the ratio of the integration
from the internal olefins and the sum of the integration from the external olefins, the
values of n and DP can be calculated. A typical spectrum of the olefin region along with
the calculation of DP is presented in Figure 2-6. With the DP, the number average
molecular (Mn) can be calculated simply by multiplying by the DP by the molecular
weight of the repeat unit (Mo). This technique is only useful for molecular weight
determination of polymers of relatively low molecular weight (40 kDa is an approximate
57
upper limit). As the molecular weight is increased, the integration from the end groups
eventually disappear and the degree of error is too large for accurate molecular weight
measurements.
R
ADMET H H H
H H H
2n H
internal olefin
2n
6
DP = 1 + n
external olefin
Figure 2-5. Relative NMR integration of internal and external olefins for NMR end-
group analysis
29.27
4.61 + 10.00
= 2.00
DP = 1 + n = 1 + 3(2.00)
DP = 7.00
Mn = MoxDP
4.61 29.27
6.00
ppm (tl)
5.50
10.00
5.00
Figure 2-6. Olefin region in the NMR for typical ADMET polymers with long methylene
spacers. The calculation to obtain the degree of polymerization shows that
this particular sample is only oligomer (DP = 7.00)
IIIII IIIIIII
CHAPTER 3
SOLID-STATE ADMET OF SEMICRYSTALLINE POLYMERS
3.1 Introduction
Although olefin metathesis has been a latecomer to the modem chemical scene, it
has become a standard chemical conversion in both organic41 and polymer42 chemistry,
particularly in the past decade. As such, it is important to understand the different
qualities and attributes of each mode of conversion. Of these different modes, acyclic
diene metathesis (ADMET) has been widely employed within our research
laboratory.60,61,62 Since the work that led to ADMET being recognized as an important
mode of olefin metathesis,121,124,125,126 a variety of unique polymer structures and
architectures have been synthesized in our laboratory. These include precisely branched
polyethylene,63'64'65'127 silicone-containing polymers,128,129,130,131,31313,134 metal-contain-
ing polymers,135'136 amino acid-containing polymers,137'138'139'140'141 branch copoly-
mers,142,143 and a number of other functionalities integrated into the polymer
backbone.144,145,146,147,148,149,150'151'152'153'154'155'156'157 Additionally work has been done to
better understand the areas where ADMET excels, as well as its limita-
tions.112,113,114,158,159,160,161 Similarly, the work herein intends to expand upon the under-
standing of this polymerization, aiming to explore the possibilities of conducting
ADMET in the solid state.
Before beginning this discussion, it is important to understand what is meant by
"solid-state polymerization." Solid-state polymerization has traditionally been divided
into two classes, crystal-to-crystal polymerization and equilibrium condensation polym-
erization.3'4'8 The process that this chapter aims to demonstrate for ADMET
polymerization is of the latter category. This type of SSP is a step-growth condensation
reaction that is conducted on semicrystalline polymers. This means that the growing
polymers have some degree of both amorphous and crystalline character, having both a
glass transition temperature (Tg) and melting transition (Tm). The polymerization is
conducted above the Tg of the polymer, which allows for enhanced molecular mobility in
the amorphous region due to a softening within these non-crystalline areas. It is
important to keep in mind that this is termed as a solid-state polymerization in industry
even though it is not strictly a crystalline solid. The amorphous region may be softened
above the Tg, however the crystalline region remains intact below the Tm, and the shape
of the particle is maintained throughout the SSP. The enhanced molecular motion within
the amorphous region facilitates the release of a small condensate molecule (ethylene, in
the case of ADMET chemistry as shown in Figure 1-24), which can be removed at the
surface of the solid to drive the increase in molecular weight. Since solid-state
condensation is widely used in industry in the syntheses of polyesters, polyamides, and
polycarbonates, many of the experiments herein were aimed at mimicking these
processes to obtain molecular weight increase in the solid state.
The one distinct advantage that ADMET has over these industrial condensation
conversions lies in the volatility of its condensate, ethylene. Removal of the condensate
can be the rate-determining factor in solid-state polycondensation. Table 1-2 compared
the condensate boiling points of typical industrial SSPs and ADMET. The industrial
examples all require temperatures of greater than 100 C to drive off the condensate,
while ethylene can be removed at much lower temperatures (BP = -104 C). This
difference allows for reactions to be performed at temperatures where metathesis
catalysts are typically active (30-60 C). Lower temperatures are also more economically
viable since less energy is required for heating.
Additional techniques were employed to further aid in condensate removal and
thereby mimic the industrial processes. In industry, the growing polymer is subjected to
flow of an inert gas or vacuum during the solid-state polymerization. In these experi-
ments, condensate removal was typically aided by the flow of argon through the reaction
vessel, although vacuum was occasionally applied where noted. Also, in some experi-
ments, the prepolymer was ground into smaller particles in a mortar and pestle to mimic
the industrial extrusion process, which is done to produce a higher amount of surface area
for the condensate to escape.
The monomer chosen for the initial work in this area is 1,9-decadiene (11 in Figure
3-1) for its simplicity in nature. This "benchmark" hydrocarbon ADMET monomer lacks
any polar or protic functionality, and thus allows for the aspects of the process to be
scrutinized rather than potential complicating effects, such as the complexation or
reaction with the catalyst. Since this monomer is a liquid, the polymerization to
polyocteneylene (12) is started in the melt state with subsequent crystallization of the
resulting polymer, and continued polymerization in the solid state. These prepolymer-to-
polymer studies are intended to be a close recreation of solid-state condensation
processes performed in industry.
Monomers with ketone and alcohol functionality were also shown to undergo
polymerization under typical ADMET conditions.152,153,154 Alcohol monomer 13 and
ketone monomers 15 and 16, which polymerize to 14, 17, and 18, respectively (Figure
3-1), were synthesized to show that solid-state ADMET could be conducted with a
variety of functionality. These monomers of higher-melting temperatures were initially
chosen in order that solid-state experiments that start from the monomer and then proceed
directly to high polymer could be attempted. There are several examples of solid-state
polycondensations that do undergo polymerization from the monomer; however, low
conversions have been typically observed.4'5'29'30 Solid mixtures of monomer and catalyst
were made by freeze-dying them together, or simply by grinding the solid monomer and
catalyst together in a mortar and pestle.
Hydrocarbon 6 ADMET
11 12
OH OH
Alcohol ADMET -
Functionalized 9 9 ADMET9 9
13 14
O 0
Ketone ADMET _
Functionalized x x ADMET nx +
15 x=9 17 x=9
16 x=8 18 x=8
0 0
Br Br
Ph Ph
Copolymer ADMET
q q 9 9n +
19 20
Figure 3-1. Monomer used in solid-state ADMET polymerization
A polystyrene graft copolymer was also selected to demonstrate solid-state
ADMET with different polymer architecture. Macromonomer 19 (Figure 3-1) is
essentially a chain of polystyrene grown from a diene by atom transfer radical
polymerization (ATRP),8'162 which could then be polymerized by ADMET.143 This type
of monomer and resulting polymer (20) would presumably be amorphous in nature since
it is essentially in a polystyrene environment.
Key factors in any solid-state polycondensation are the nature and the lifetime of
the catalyst employed, in this case, metathesis catalysts. Three of the most widely used
metathesis catalysts are the Grubbs first- (5)98 and second- (6)103 generation catalysts and
Schrock's molybdenum catalyst (2).83 Experiments were conducted comparing the
activity of these catalysts for a variety of conditions. This chapter demonstrates that the
two Grubbs catalysts do indeed promote solid-state ADMET polymerization.
3.2 Experimental
3.2.1 General
All reactions were conducted in flame-dried glassware under an argon atmosphere
unless otherwise noted. Reagents were purchased from Fischer or Aldrich and were used
as received unless otherwise noted. 1,9-Decadiene was distilled from CaH2 before use.
Metathesis catalysts used in this chapter, first- (5)98 and second- (6)103 generation Grubbs
and Schrock's molybdenum (2),83 were synthesized by known methods.
Instrumentation for GPC and NMR are discussed in Chapter 2. Thermal analysis
was conducted on a Perkin-Elmer DSC 7 with a Perkin-Elmer CCA 7 temperature
regulator. Gas Chromatography was performed on a Shimadzu GC-17A chromatograph
equipped with a RTX-5 (Restek Corp.) 15 m column and a HP-5 (Hewlett Packard) 25 m
column with FID detection. Elemental analyses were performed by Atlantic Microlab,
Inc (Norcross, GA). Mass spectrometry (MS) was conducted by the University of
Florida Mass Spectrometry Services. Melting points were performed on a Laboratory
Devices Mel-Temp apparatus.
3.2.2 Monomers Syntheses
3.2.2.1 Synthesis of tricosa-1,22-dien-12-ol (13)
This monomer was synthesized by one of two Grignard reactions, one using the
alkenyl bromide and the other using the alkenyl chloride. Physical attributes of the
compound from both routes matched after purification.
For the bromide route, a procedure from the literature was modified.154 Magnesium
turnings (2.12 g, 87.2 mol) were ground in a mortar and added to a three-neck round-
bottom flask equipped with a condenser and an addition funnel. A solution of 11-bromo-
undec-1-ene (22) (16 mL, 17 g, 0.074 mol) in dry diethyl ether (60 mL) was added
dropwise while warming gently in a water bath (up to 40 OC), when a greenish black
turbid solution formed. After complete addition (1.5 h), the solution was refluxed for an
additional 2 h, at which time a solution of ethyl format (2.5 mL, 2.3 g,
0.031 mol) in dry ether (10 mL) was added dropwise (15 min). The resulting solution
was refluxed overnight. The reaction mixture was poured onto 200 mL of
3 N HCl/ice, causing a white solid to precipitate out. The layers were separated, and the
aqueous portion was extracted with ether (30 mL). The combined organic layers were
washed twice with 3 N HC1 (50 mL), twice with water (50 mL), and dried over MgSO4.
Upon evaporation of solvent, 10.82 g of the crude white product was obtained. This
material was purified by flash chromatography using 40 % DCM in hexanes, followed by
100 % DCM to yield 6.33 g (61 %) of white tricosa-1,22-dien-12-ol (13) (MP =
54-57 C). Purification by 10 % ethyl acetate in hexanes has also been reported.154 1H
NMR (CDC13, ppm): 6 = 5.89-5.73 (m, 2H), 5.04-4.89 (m, 4H), 3.69-3.52 (br, 1H), 2.09-
1.98 (q, 4H), 1.62-1.52 (br s, 1H), 1.50-1.18 (br m, 32H). 13C NMR (CDC13, ppm): 6 =
139.47, 114.31, 72.24, 37.71, 34.03, 29.92, 29.83, 29.76, 29.70, 29.35, 29.15. 25.87.
ESI/FT-ICR HRMS: Calcd for C23H440Na ([M+Na] ): 359.3290; Found: 359.3295.
Elemental Anal. calcd for C23H440: C 82.07, H 13.18, O 4.75; Found: C 82.14, H 13.26,
O 4.79.
For the chloride route, magnesium turnings (4.11 g, 87.2 mol) were ground in a
mortar and added to a similar setup as above. A solution of 11-chloro-undec-l-ene (23)
(23.6 g, 0.125 mol) in dry THF (125 mL) was added dropwise (1 h) and the solution
allowed to stir for an additional hour. The temperature of the reaction flask was reduced
to 78 C with a dry ice/acetone bath and ethyl format (4.6 mL, 4.2 g, 0.057 mol) was
added dropwise, after which the resultant solution was allowed to reflux overnight. The
reaction was quenched with 3 N HC1 and the aqueous portion was extracted with ether.
The combined organic layers were washed twice with 3 N HC1, twice with water, and
dried over MgSO4. After evaporation of the solvent, 10.8 g (96 %) of the crude white
product was obtained the resultant crude product was recrystallized in acetone twice to
obtain white tricosa-1,22-dien-12-ol (13), having the same physical properties as that
described above.
3.2.2.2 Synthesis of tricosa-1,22-dien-12-one (15)
This synthesis was performed by Paul van Gervan within our research group. A
slurry of pyridinium chlorochromate (2.69 g, 12.5 mmol) in 10 mL of dry
dichloromethane (DCM) was prepared in a 100 mL three-neck round-bottom flask
equipped with two stoppers and an addition funnel, to which a solution of diene alcohol
17 (3.26 g, 9.69 mmol) in 20 mL of dry DCM was added dropwise. After the addition
(2 h), the black slurry was stirred overnight at room temperature. Diethyl ether
(25 mL) was added and the resulting turbid solution was decanted. This addition of ether
and decanting was repeated twice. The ether/DCM solution was passed over a plug of
silica and rotovaped to yield a white solid with a green tint. The solid was dissolved in
35 mL of hexanes, washed three times with 25 mL of water, dried over MgSO4, and
subsequently evaporated under reduced pressure to yield 2.62 g (80.1 %) of white tricosa-
1,22-dien-12-one (15) (MP = 58-59 C). 1H NMR (CDC13, ppm): 6 = 5.90-5.73 (m, 2H),
5.04-4.88 (m, 4H), 2.38 (t, 4H, J = 7.2 Hz), 2.03 (q, 4H, J = 7.2 Hz), 1.55 (p, 4H, J =
7.2 Hz), 1.44-1.17 (br, 24H). 13C NMR (CDC13, ppm): 6 = 211.96, 139.44, 114.33,
43.04, 34.02, 29.62, 29.48, 29.32, 29.13, 24.01. HRMS: Calcd for C23H420: 334.3236;
Found: 334.3240. Elemental Anal. calcd for C23H420: C 82.57, H 12.65; Found: C 80.88,
H 12.59.
3.2.2.3 Synthesis of heneicosa-1,20-dien-1l-one (16)
Synthesis was performed using a procedure published in the literature,139 which was
modified from the original synthesis by Zantour and coworkers.163 10-Undecenoyl
chloride (21.5 mL, 0.100 mol, 20.3 g) and dry diethyl ether (150 mL) were added to a
500 mL three-neck round-bottom flask equipped with a reflux condenser and an addition
funnel, and then cooled to 0 OC. Triethyl amine (25.1 mL, 0.180 mol, 18.2 g) was added
dropwise, instantly precipitating white triethylammonium chloride salt. The reaction
mixture was allowed to warm to room temperature and stir for 24 h. After salt filtration
and ether evaporation, the intermediate lactone was mixed with 100 mL of a
2.0 M solution of NaOH (8.00 g, 0.20 mol) and refluxed overnight under argon. The
reaction mixture was acidified with 200 mL of 1.0 M HC1 and extracted three times with
ether. The combined ether layers were washed twice with 1.0 M HC1, twice with
saturated NaHCO3, and once with brine, dried over MgSO4, and recrystallized in MeOH
to yield 9.60 g (62.7 %) of pure heneicosa-l,20-dien-11-one (16) (MP = 52-55 OC). 1H
NMR (CDC13, ppm): 6 = 5.90-5.72 (m, 2H), 5.04-4.87 (m, 4H), 2.38 (t, 4H, J = 7.2 Hz),
2.02 (q, 4H, J = 7.2 Hz), 1.56 (m, 4H), 1.35 (m, 4H), 1.32-1.20 (br, 16H). 13C NMR
(CDC13, ppm): 6 = 211.93, 139.42, 114.35, 43.03, 34.01, 29.58, 29.52, 26.47, 29.29,
29.11, 24.09. HRMS: Calcd for C21H380: 306.2923; Found: 306.2932. Elemental Anal.
calcd for C21H380: C 82.28, H 12.50, 0 5.22; Found: C 81.96, H 12.74, 0 4.99.
3.2.2.4 Synthesis of the polystyrene-graft macromonomer (19)
This monomer was prepared via the literature method.143 The atom transfer radical
polymerization (ATRP) step is reported here while precursor steps to achieve initiator 24
are reported therein. Styrene was washed with NaHCO3 and deionized water to remove
inhibitors and distilled over CaH2 under reduced pressure before use. Diphenyl ether was
recrystallized in 95 % ethanol in water, melted, dried over CaC12, and then fractionally
distilled before use. Copper(I) bromide (0.43 g, 3.0 mmol) and 2,2'-bipyridine (0.94 g,
6.0 mmol) were weighed out and placed into a flame-dried, argon-purged, 50-mL
Schlenk tube with a stir bar. Diphenyl ether (5 mL) was added via a syringe, and the
mixture was subsequently degassed with three freeze-pump-thaw cycles. Initiator 24
(0.75 g, 1.5 mmol) and styrene (9.37 g, 90 mmol) were added via a syringe to the
aforementioned solution. The resultant mixture was stirred for 3 h at 110 OC while
maintaining a slight argon flow. The solution changed from brown to green upon
heating. The reaction mixture was cooled to room temperature and quenched with THF
(60 mL), filtered through a neutral bed of alumina to remove residual catalyst, and filtrate
was concentrated to 30 mL. Macromonomer 19 was obtained by precipitation into cold
methanol. Finally, this solution was filtered, and the white solid product 19 was dried
under reduced pressure. Integrations in NMR showed that approximately 60-65
polystyrene units were on each chain. 1H NMR (CDC13, ppm): 6 = 7.23-6.870 (br,
203H), 6.87-6.28 (br, 130H), 5.90-5.70 (m, 2H), 5.08-4.90 (m, 4H), 4.50 (t, 1H), 3.51 (d,
dH), 2.04 (q, 12H), 1.98-1.68 (br, 56H), 1.68-1.08 (br m, 176H), 1.02-0.80 (br m, 8H).
13C NMR (CDC13, ppm): 6 = 177.81, 157.16, 145.24, 139.14, 129.66, 127.91, 125.60,
123.13, 118.82, 114.12, 66.81, 41.82, 40.33, 37.02, 33.80, 31.26, 29.96, 29.59, 29.14,
28.93, 26.67. NMR end-group analysis: Mn = 7400. GPC: Mn = 6800, Mw = 9200, PDI
= 1.35.
3.2.3 Polymer Syntheses
Methods of creating a solid solution, running a solid-state reaction, modes of
quenching, and types of molecular weight characterization are discussed in Chapter 2.
The details of what was used for each particular experiment including catalyst loading
and reaction temperature are specified in the results and discussion section as each
experiment is conveyed. In a typical experiment, each sample was analyzed for
molecular weight by GPC and/or NMR end-group analysis. Any additional character-
ization will appear in the results and discussion as well. Full characterization has not
been included since the polymers synthesized herein have been synthesized before, and
only trends in molecular weight increase in the solid state were desired.
3.2.4 Other Molecules Synthesized
These compounds (21-23) are all known and the physical properties of the product
match those of Aldrich and have not been included. These procedures appear in the
literature from Baughman and coworkers164 within our group.
3.2.4.1 Synthesis of undec-10-en-l-ol (21)
Zinc undecylenate (250 g, 0.579 mol) was added over 30 min via a powder funnel
to a stirred slurry of lithium aluminum hydride (LAH) (25.0 g, 0.659 mol) in dry THF
(400 mL) in a 5 L round bottom flask at 0C. After the addition, the solution was
allowed to warm to room temperature over 1 h while stirring. The reaction was quenched
via addition of deionized water (25 mL), 15 % (w/v) NaOH (25 mL), and more water (75
mL) waiting approximately 15 min between additions. The solution was allowed to stir
until cool, and the reaction mix appeared as a white slurry. All precipitate was filtered,
and the solution was concentrated to a turbid oil. The crude mixture was dissolved in
ether and stirred over MgSO4 (15 g) for 45 min. The solution was filtered and
concentrated to afford 85.8 g (87 %) of undec-10-en-l-ol (21) as a colorless oil (97 %
pure by GC).
3.2.4.2 Synthesis of 11-bromo-undec-l-ene (22)
A solution of undec-10-en-l-ol (21) (150 g, 0.881 mol) and carbon tetratbromide
(323 g, 0.976 mol) in dichloromethane (800 mL) was prepared in a 2 L flask and cooled
to 0 OC. Triphenyl phosphine (255 g, 0.976 mol) was added via a powder funnel in
portions over 30 min with vigorous stirring. Upon addition of the phosphine, the
colorless solution turned a pale brown color and was stirred for an additional 2 h at room
temperature. The mixture was concentrated to a brown oil and quickly added to stirring
hexanes (1 L). The white precipitate was filtered, and the remaining solution was
concentrated and fractionally distilled yielding 195 g (95 %) of 11-bromo-undec-l-ene
(22) as a colorless liquid. In addition to Aldrich, the physical properties matched those
published previously in our group.154
3.2.4.3 Synthesis of 11-chloro-undec-1-ene (23)
In an argon-purged 3 L round-bottom flask, distilled thionyl chloride (259 g,
2.17 mol) was added dropwise over 1 h via cannula into a solution of undec-10-en-l-ol
(21) (200 g, 1.28 mol) in pyridine (50 mL). Upon complete addition, the reaction was
heated to 50 OC for 2 h, cooled, and quenched via the addition of water
(300 mL) and diethyl ether (300 mL) letting stir for 1 h. The remaining mixture was
extracted, and the organic phase was washed twice with saturated NaHCO3 (150 mL) and
once with distilled water (100 mL). The solution was dried over MgSO4, concentrated,
and vacuum distilled to yield 169 g (70 %) of 11-chloro-undec-l-ene (23) as a colorless
oil.
3.3 Results and Discussion
Much of the data that appears in this section also appears in the literature;165
however, several experiments that do not appear in the original communication have been
included, along with a more thorough discussion of the results.
3.3.1 Prepolymer-to-Polymer Experiments
Solid-state experiments that start with partially polymerized material closely
mimic industrial solid-state polycondensation. Monomers of varying functionality and
architecture (Figure 3-1) were shown to undergo solid-state ADMET in a manner similar
to the SSPs conducted in industry.
3.3.1.1 Solid-state polymerization of 1,9-decadiene
This initial work with this simple hydrocarbon monomer demonstrated that
ADMET polymerization could be conducted in the solid state. It was found that by
starting the polymerization in the liquid state at a constant temperature, the growing
polymer would crystallize, and then continue molecular weight increase in the solid state.
Data for catalysts 5 and 6 at 500:1 monomer to catalyst ratios is displayed in Table 3-1
for 30 and 45 C.
Table 3-1. Results of solid-state polymerizations of monomer 11 with periodic sampling
Ct Mon: Cat Temp Number Average Molecular Weights"
Ratio (oC) at Specified Reaction Times
5 5001 30 3 h 6h 12 h 24 h 48 h 72 h 96 h 168 h
2,800 3,700 6,600 9,400 9,900 19,000 20,000 21,000
5 500 1 45 3 h 6h 12 h 30 h 48 h 72 h 107 h
5,400 6,400 7,200 16,000 30,000 30,000 31,000
6 500 1 30 3 h 6h 12 h 24 h 48 h 119h
2,200 5,800 7,000 7,800 7,900 8,800
6 500 1 4 3 h 6h 12 h 24 h 48 h 72 h 96 h 168 h
8,900 12,000 12,000 15,000 18,000 18,000 21,000 21,000
SMn determined by GPC in chloroform relative to polystyrene standards.
These results demonstrate unequivocally that solid-state polycondensation is
occurring at temperatures as low as 30 oC, with the possibility of the reaction occurring at
even lower temperatures. This is the first example of a solid-state polycondensation
being conducted as low as room temperature, as all other examples are conducted at
elevated temperatures. Figure 3-2 shows the GPC traces for polymerizations done with
first-generation Grubbs catalyst at 45 oC. As the polymerization proceeds, the number
average molecular weight increases from 5,400 to 31,000 g/mol, all while in the solid
state. It should be noted that all the molecular weights reported herein are well within the
range normally expected for equilibrium step-growth polymerization chemistry. Table
3-2 illustrates that the molecular weights observed in this work are within the range of
useful molecular weights for typical step growth polymers, which exhibit useful
properties with number average molecular weights of less than 25,000.166
71
31,000
107 hrs
72 hrs
48 hrs
16,000 3 h,
S30 hrs
12 hrs
6 hrs
5,400
3 hrs
0 5 10 15 20
Retention Time (min)
Figure 3-2. Gel Permeation Chromatography traces of the SSP of 1,9-decadiene with
first-generation Grubbs at 45 C sampled periodically (The molecular weights
listed at 3, 30, and 107 h are number average molecular weights determined
by GPC in chloroform relative to polystyrene standards.)
Table 3-2. Useful molecular weights of typical step-growth polymers
Polymer Useful Molecular Weights
Polyesters: Mn ~ 15,000 to 25,000
Polyamides: Mn ~ 15,000 to 30,000
Polycarbonates: Mn ~ 12,000 to 30,000
ADMET Polymers: Mn ~ 10,000 to 30,000
Both catalysts 5 and 6 promote solid-state ADMET polymerization. First-
generation Grubbs (5) was found to be the more active of the two, largely due to its
higher reactivity at lower temperatures, reaching number average molecular weights of
30,000 g/mol in 3 d at 45 C. Second-generation Grubbs (6) is slower, but seems to
continue to promote polymerization at later stages of the reaction. This observation may
be the result of the longer lifetime of complex 6 as compared with the first-generation
catalyst. This is similar to what is seen in the literature in solution. The half-life of
second-generation Grubbs is found to be 6 h at 55 C in benzene,108 compared to 40 min
for first-generation Grubbs;103 however, we suspect that the lifetimes of these complexes
are longer in the solid state compared to those in solution.
Experiments with Schrock's molybdenum catalyst (2) were conducted at 45 C
with a 1500:1 monomer to catalyst ratio. Since this catalyst is much more sensitive to
oxygen, the reaction mixture had to be started in a glove box under an argon atmosphere
and then transferred to an outside line to achieve argon flow. Two experiments were run
side by side for 75 h, one with sampling in the glove box, and one without. For both
experiments, the molecular weight of the first sampling period equaled the molecular
weight of the final sampling at 75 h (Table 3-3). Any molecular weight fluctuation
observed is within the error of the technique. The likely reason that Schrock's catalyst
was not observed to be active is decomposition of the catalyst. Although two
experiments cannot rule out the possibility that catalyst 2 can be active in the solid state,
further work herein was focused on the activity of the Grubbs catalysts 5 and 6 in the
solid state.
Table 3-3. Results of solid-state polymerizations using Schrock's catalyst 2 and
monomer 11
Method of Mon: Cat Temp Number Average Molecular Weights"
Sampling Ratio (C) at Specified Reaction Times
Outside 15001 45 3 h 25 h 50 h 75 h
Glove box 1,600 1,500 1,500 1,600
Inside 15001 45 8 h 27 h 75 h
Glove box 1,800 1,800 1,800
a Mn determined by GPC in chloroform relative to polystyrene standards.
Since this is a solid-state reaction where molecular mobility is relatively low, it was
thought that increasing the catalyst loading would potentially provide a route to higher
conversion. The catalyst loading was increased five-fold (100:1 monomer to catalyst
ratio) with the second-generation catalyst (6) and is compared with the corresponding
500:1 reactions at 30 and 45 C in Table 3-4. At both temperatures, comparable
molecular weights were achieved with both catalyst loadings, although, the observed
molecular weights of the experiments with higher catalyst loadings actually tended to be
slightly lower. These experiments suggest that either catalyst loading, or anywhere in
between, is acceptable to achieve molecular weight increase in the solid state, but higher
catalyst loadings do not promote the polymerization to higher conversion.
Table 3-4. Catalyst loading comparison using monomer 11 and catalyst 6
C Mon: Cat Temp Number Average Molecular Weights"
Ratio (C) at Specified Reaction Times
6 500 1 30 3 h 6h 12 h 24 h 48 h 119h
2,200 5,800 7,000 7,800 7,900 8,800
6 1001 30 5 h 8 h 24 h 48 h 72 h 120 h
2,800 5,800 6,700 6,800 6,700 8,100
6 5001 45 3 h 6h 12 h 24 h 48 h 72 h 96 h 168 h
8,900 12,000 12,000 15,000 18,000 18,000 21,000 21,000
6 1001 45 3 h 6h 24 h 48 h 72 h 96 h 168 h
6 100 : 1 45
4,000 5,900 10,000 14,000 15,000 16,000 18,000
SMn determined by GPC in chloroform relative to polystyrene standards.
When comparing polymerizations of the same catalyst at different temperatures,
higher temperatures achieved higher molecular weights (Table 3-1). For first-generation
Grubbs (5) at 30 OC, a number average molecular weight of 19,000 g/mol was achieved
in 3 d, while at 45 C, it reached 30,000 g/mol in the same amount of time. A similar
effect was found with second-generation Grubbs catalyst (6). At 30 C, the number
average molecular weight does not exceed 10,000 g/mol in 5 d, while reaching
approximately 21,000 g/mol at 45 C in the same amount of time.
Similar experiments to these were conducted at 60 OC where ultra-high molecular
weights were achieved with catalyst 6. However, all but one of these experiments did not
appear to be a solid-state reaction as the reaction mixtures were viscous melts. The
presence of liquid properties was tested simply by making a hole in the viscous mixture
with a spatula; hours later the hole would be filled because the material had exhibited
flow properties. In one experiment (line 1 in Table 3-5), however, the reaction mixture
did indeed solidified before the first sampling time at 3 h. Several attempts to reproduce
this experiment and decipher why this solidification occurred were unsuccessful. All
experiments conducted at 60 OC are presented in Table 3-5.
The most striking piece of data from these experiments is that the molecular weight
achieved by the one solid-state reaction (line 1) was 122,000 g/mol after four days. After
7 d, however, the molecular weight decreased to 79,000 g/mol, which is what is observed
in typical solid-state polycondensation with extended reaction periods (Figure 1-11).
Since a prior attempts at 60 C started outside the glove box with catalysts 6 and 5 (line 2
and 5, respectively) had been observed to be a viscous melt, this solidification was
initially thought to be from less catalyst decomposition through starting the reaction
inside the glove box. To test this theory, another experiment starting in the glove box
(line 3) was conducted in the exact same manner to yield a viscous melt during
polymerization. This suggested that it was not lack of catalyst decomposition that caused
it to solidify. However, ultra high molecular weight was observed in this experiment as
well (131,000 g/mol after 7 d). This high molecular weight is more easily reasoned for a
melt reaction. These polymerizations are neat reactions (no solvent), therefore when in
the melt state, the olefinic species, both monomer and growing polymer, are at the
Table 3-5. Experiments conducted at 60 OC using monomer 11
Line Cat Mon: Cat Started in Number Average Molecular Weights" State
R itiin GrlnvP Rnx it Snecrifird Renctinn Timoe Ohervdrl
1 6 500: 1
2 6 500: 1
3 6 500: 1
4 6 100: 1
5 5 500: 1
3h
yes
12,000
3h
no
17,000
3h
yes
14,000
3h
no
15,000
3h
no
12,000
6h
11,000
6h
22,000
6h
24,000
6h
20,000
6h
15,000
12 h 24 h 48 h 72 h 96 h 168 h
17,000 16,000 14,000 37,000 122,000 79,000
12 h 24 h 48 h 72 h 96 h 168 h
44,000 47,000 66,000 80,000 80,000 98,000
12 h 24 h 48 h 72 h 96 h 168 h
23,000 26,000 37,000 43,000 52,000 131,000
12 h 24 h 48 h 72 h 96 h 168 h
20,000 29,000 28,000 35,000 34,000 35,000
12 h 24 h
14,000 15,000
48 h 72 h
14,000 14,000
96 h 168 h
13,000 15,000
Solidb
Melt
Melt
Melt
Melt
a MA determined by GPC in chloroform (line 1-3, 5) or THF (line 4) relative to polystyrene standards. b This state was not
reproducibly observed.
highest concentration possible. This, combined with the molecular mobility of a liquid
state, gives the olefin end-groups the best chance to find each other and achieve this ultra-
high molecular weight. Another possibility that was tested was that the solidification
came from an error in weighing out the catalyst causing a higher catalyst loading. An
experiment (line 4) with a higher catalyst loading (100:1) was run, again yielding a
viscous melt. Although this is not a solid-state reaction, it is interesting to note that this
higher catalyst loading did indeed yield significantly lower molecular weights. Even
though the cause of the solidification (line 1) is not clear, the fact remains that it did
occur, and the molecular weight did increase from 12,000 g/mol to 122,000 g/mol in the
solid state. This is a 10-fold increase to yield ultra high molecular weight polymer.
The condensate in solid-state polycondensations is removed in industry by a flow
of inert gas or by vacuum, whereby the choice is usually based on convenience. For the
majority of these experiments, flow of argon through the reaction vessel was used.
However, the use of vacuum to remove the condensate was tested. Table 3-6 compares
the number average molecular weights of these two methods at 45 C with catalyst 6.
For the vacuum experiment, a round-bottom flask was equipped with a bath condenser,
and a Schlenk adapter inside a glove box. The monomer and catalyst were added to the
flask in the glove box, and then the apparatus was transferred to an outside line with the
Schlenk stopcock closed. After filling the bath condenser with dry ice/isopropanol and
submerging the reaction flask at 45 C, the apparatus was subjected to intermittent
vacuum until the reaction mixture solidified. Sampling for this experiment was con-
ducted inside the glove box. The molecular weight achieved at the end of the vacuum
experiment was less than half that of the argon experiment. Because of this observation
and complications with volatility of monomer 11, argon flow is the method of choice for
these experiments.
Table 3-6. Comparison of argon and vacuum experiments using monomer 11 and
catalyst 6
Condensate Mon: Cat Temp Number Average Molecular Weights"
Removal Ratio (0C) at Specified Reaction Times
argon 5003 h 6 h 12 h 30 h 48 h 72 h 107 h
argon 500 : 1 45
5,400 6,400 7,200 16,000 30,000 30,000 31,000
vacuum 5001 45 5 h 27 h 72 h 168 h
vacuum 500 : 1 45
2,300 12,000 13,000 13,000
a Mn determined by GPC in chloroform (argon) or THF (vacuum) relative to polystyrene
standards.
Experiments also were conducted to increase the surface area of the solid mixture
in order to facilitate escape of the condensate from the surface of the solid. This also
further simulates the industrial process whereby the prepolymer is extruded into small
pellets for the same reason. One such experiment is illustrated in Figure 3-3. 1,9-Dec-
adiene (11) was polymerized in the melt with second-generation Grubbs at 70 OC for 10
min and then quickly crystallized by cooling to room temperature. The resulting low
molecular weight semicrystalline polymer (Mn = 1,300 g/mol by NMR) was
6 6
10 min at 70 C 6 7 Days at 45 C
(melt) (solid state)
11 Mn= Mn
1,300 g/mol 23,000 g/mol
Figure 3-3. Prepolymerization in the melt of 1,9-decadiene, followed by polymerization
in the solid state using second-generation Grubbs ruthenium catalyst (6)
(Number average molecular weights (Mn) were obtained by NMR end-group
analysis in CDC13.)
ground into a fine powder with a chilled mortar and pestle. This finely divided solid was
brought to 45 OC and sampled over time as before, showing that Mn steadily increases
over the course of a week, reaching 23,000 g/mol by NMR. This represents an
approximate 20-fold increase in molecular weight, all while in the solid state.
3.3.1.2 Alcohol and ketone functionalized monomers
Alcohol monomer 13 and ketone monomers 15 and 16 were synthesized in the
routes designated in Figures 3-4 and 3-5. Alcohol 13 was synthesized via one of two
Grignard reactions from alkenyl bromide 22 or alkenyl chloride 23 (Figure 3-4).154 These
alkenyl halides metathesis synthons, as well as alkenyl alcohol synthon 21, were
synthesized via routes laid out within our group from readily available, inexpensive
starting materials.164 Ketone 15 can be easily made by oxidation of alcohol 13 with
pyridinium chlorochromate (PCC) (Figure 3-4). Ketone 16 was synthesized via a
completely different route from the acyl chloride (Figure 3-5). This synthesis was
LAH OH CBr4,PPh3 Br
20 Zn THF DCM
21 22
SOCI2 O 0
pyridine Mg H OEt
H OEt
o OH 0
9PCC
09 O9 DCM 9 9
23 H OEt 13 15
Figure 3-4. Synthetic route to alcohol 13 and ketone 15
modified within our group139 from a procedure first performed by Zantour and
coworkers.163 This procedure is a one-pot synthesis from 10-undecenoyl chloride, which
happens to be only acyl chloride a-olefin that is currently readily available and
inexpensive. Since ketones 15 and 16 only differ in one methylene group on each side of
the diene, they were considered equivalent in terms of these solid-state experiments.
Monomer 16 became the monomer of choice for the majority of the ketone experiments
because of simplicity of the one-pot synthesis. In fact, ketone 15 was only used in the
monomer-to-polymer experiments (Section 3.3.2).
O
II
C
2 NE 2
C 801 7 7
NaOH, H20
O OH 0 OH
taut. .
7 7> 7 7>-
A CO2
taut.
7 O7
OH 16
Figure 3-5. Mechanism of synthesis of ketone 16 (This mechanism has been published in
references [139] and [163].)
Experiments with monomers 13 and 16 similar to those described above for
1,9-decadiene demonstrated that a variety of functionalities are amenable to solid-state
ADMET polycondensation. For the Grubbs catalysts 5 and 6, functionality of this type is
not an issue since ruthenium reacts preferentially with olefins before alcohol and ketone
functionalities (Table 1-1).85 Since higher molecular weight increase can be achieved
with increasing surface area, as was observed with 1,9-decadiene in Figure 3-3, all of
these alcohol and ketone experiments that start from a prepolymer were ground in a
mortar and pestle in air to yield a fine powder before beginning the solid-state portion of
the experiment. The melting points of polymers 14 and 18 was significantly higher than
that of polymer 12, so the grinding could be done at room temperature (not with a chiller
mortar). Another difference from the 1,9-decadiene experiments is that these experi-
ments were conducted partially in the melt state at elevated temperatures to make a solid
solution since these monomers were solids at room temperature. The prepolymer
synthesized in the melt portion of the reaction has been labeled as the "start" sample, as
all subsequent molecular weight increase was in the solid state. Examples of these
experiments are presented in Table 3-7 for alcohol 13 and Table 3-8 for ketone 16.
Table 3-7. Examples of solid-state polymerizations of alcohol monomer 13
Cat Mon: Cat Melt Reaction SSP Number Average Molecular Weights"
Ratio Time Temp Temp at Specified Reaction Times
6 300 : 1 5 min 80 C 60 C Start Id 4d 7d
2,100 5,200 5,600 6,500
6 300:1 5 min 115 C 60 C Start 4d 7d
8,200 9,200 9,900
SMn determined by NMR end-group analysis in toluene-ds at 80 OC.
The two examples of polymerizations from alcohol 13 that appear in Table 3-7
have different melt reaction temperatures. At 80 C, a number average molecular weight
of 2,100 g/mol by NMR end-group analysis was achieved in 5 min. During seven days of
reacting in the solid state, the molecular weight increases to 6,500 g/mol, a three-fold
increase while in the solid state. At 115 C, a higher molecular weight of 8,200 g/mol
was achieved in the sample amount of time in the melt phase. Although steady increase
in molecular weight is observed, the change in molecular weight in the solid state to
9,900 g/mol after 7 d was less significant. This may be due to catalyst decomposition at
the higher reaction temperatures during the melt portion of the reaction.
For the five ketone experiments in Table 3-8, the higher temperature (110-115 C)
were actually required to maintain the polymer as viscous liquid during the melt portion
of the reaction. All of these examples showed a trend of molecular weight increase in the
solid state. Using the second-generation catalyst (6), a molecular weight of 8,700 g/mol
was observed by GPC after 3 min at 110 OC. The molecular weight then increased to
18,000 g/mol over the next seven days, more than doubling while in the solid state. At 10
min in melt reaction, a similar trend was observed, increasing from 6,500 g/mol to 11,000
g/mol while in the solid state. Unlike the alcohol (13), the longer time in the melt did not
yield the higher molecular weight. However it did lead to a slightly less significant
conversion as was observed in Table 3-7.
Table 3-8. Examples of solid-state polymerizations of ketone monomer 16
SMon: Cat Melt Reaction SSP Number Average Molecular Weights"
Ratio Time Temp Temp at Specified Reaction Times
6 5001 3 min 110C 80C Start 3 h 3 d 5 d 7 d
6 500:1 3 min 110 C 80 "C
8,700 12,000 14,000 18,000 18,000
6 500:1 10 mi 115 OC 80C Start 6h d 3d 5d 7d
6,500 9,000 9,700 9,500 10,000 11,000
5 3001 5 min 110C 80C Start 3h Id 3d 5d 7d
5 300:1 5 min 110 C 80 "C
1,100 1,400 3,000 4,600 4,700 5,300
5 3001 10 min 110C 80C Start 1 d 3 d 5 d 7 d
5 300:1 10 min 110 C 80 "C
1,100 3,200 4,400 5,200 4,900
5 300:1 45 min 115 C 80 C Start 3"d 5d 7d
1,300 2,600 3,400 3,500
a Mn determined by GPC in chloroform relative to polystyrene standards for experiments
with catalyst 6, and by NMR end-group analysis in CDC13 for experiments with catalyst
5.
Polymerizations with catalyst 5 did not achieve as high molecular weights as with
catalyst 6, although the same trends of molecular weight increase were also observed.
All three experiments with catalyst 5 shown in Figure 3-8 achieved only low number
average molecular weights of 1,100-1,300 g/mol by NMR end-group analysis during the
melt portion of the experiment. The most significant molecular weight increase here was
again the reaction that spent the shortest time in the melt phase. At 5 min in the melt, the
ending molecular weight after 7 d was 5,300 g/mol. This number got progressively
smaller with more time in the melt state reaching only 4,900 g/mol for 10 min and
3,500 g/mol after 45 min. It would appear that the longer the mixture is kept in the melt,
the more that the catalyst decomposes. Therefore, there is less catalyst left in the solid
solution for the polymerization in solid state, and a lower molecular weight is achieved.
The last entry in Table 3-8 with an extended reaction period in the melt (45 min) was
actually conducted in hopes of achieving a higher molecular weight of the prepolymer.
Instead, even after this extended reaction period, the molecular weight remained quite
low (1,300 g/mol), and it consequently had this negative impact on molecular weight
increase in the solid state due to extensive catalyst decomposition.
When comparing reactions with catalysts 5 and 6 in Table 3-8, there are two
differences to note. The mode of molecular weight determination is different, as well as
a slight difference in catalyst loading. Although these differences are present, they do not
invalidate a comparison of the results. The second-generation Grubbs catalyst performed
better in these experiments since it could better withstand the higher temperatures during
the melt portion of the reaction, and therefore more catalyst remained active for the solid-
state portion of the experiment.
Experiments were conducted with purified prepolymer as well (Table 3-9). After a
short reaction in the melt state, cooling, and grinding into a fine powder, the portion of
the prepolymer was quenched with EVE, dissolved in THF, and then precipitated in cold
methanol. After being dried under high vacuum overnight, the resulting white polymer
was reground together with fresh catalyst (300:1 repeat unit to catalyst ratio) and run
aside the typical, unprecipitated material containing the original catalyst left over from
the melt portion of the reaction. The amount of fresh catalyst used in the regrinding was
not based on the moles of polymer, but rather the moles of repeat unit within the polymer,
calculated from the weight of the solid prepolymer and the number average molecular
weight obtained from NMR end-group analysis. This allows for a more accurate
comparison between the precipitated and unprecipitated control experiments since the
original catalyst ratios for the melt reactions were based on moles of monomer. As in
any polymer precipitation, the purified product was somewhat fractionated by the
washing away of lower molecular weight portions to yield slightly higher number
average molecular weights. In both sets of experiments in Table 3-9, one for 5 min in the
melt and one for 10 min, this apparent increase in the number average molecular weight
during precipitation process was from 1,100 g/mol to 1,500 g/mol.
Table 3-9. Precipitation effects on the solid-state polymerizations with catalyst 5 and
ketone 16
a Mon: Cat Melt Reaction SSP Number Average Molecular Weightsb
Ratio Time Temp Temp at Specified Reaction Times
no 300:1 5min 110 80C Start 3h Id 3d 5d 7d
1,100 1,400 3,000 4,600 4,700 5,300
Start 7d
yes 300:1 5min 110C 80C tart
1,500 9,600
no 300:1 10 min 110 C 80 C Start Id 3d 5d 7d
1,100 3,200 4,400 5,200 4,900
Start 1 d 3 d 5 d 7 d
yes 300:1 10 min 110 C 80C tartd 3d 7d
1,500 4,700 6,300 7,000 7,500
aPrecipitation of the prepolymer after the melt portion of the polymerization. Mn
determined by NMR end-group analysis in CDCl3.
Not only were trends in molecular weight increase observed for the precipitated
experiments, but also this increase was more pronounced than in the unprecipitated
control experiments. At 5 min in the melt, a precipitated portion, reground with fresh
catalyst and allowed to polymerize for 7 d at 80 OC in the solid state, achieved molecular
weight of 9,600 g/mol. This is compared to the 5,300 g/mol achieved in the
|
|
PAGE 1
SOLID-STATE OLEFIN METATHESIS By GARRETT W. OAKLEY A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2004
PAGE 2
Copyright 2004 by Garrett W. Oakley
PAGE 3
This dissertation is dedicated to my family.
PAGE 4
iv ACKNOWLEDGMENTS First and foremost, I thank Jennifer Sippel/O akley, the girl who became my wife on October 2, 2004. I met her when I visited the Un iversity of Florida on my recruiting trip, we started dating the day I moved to Florid a, and we are getting married the same semester that I plan to graduate. She has been there throughout my entire graduate career, and the motivation, understanding, a nd love she has given me during the completion of this work is irreplaceable. I would like to thank my family, both imme diate and extended, past and present, to whom this document is dedicated, for being an endless source of support in my education and life. Without this, I would have been lost In particular, I want to thank my father, David Oakley, for all his help and advice. Ev en after eight years of schooling and going into debt, my father has encouraged me to get the best education I could. I thank my mother, Jeanne Qualey, for the time in recent y ears that we were able to spend together since she moved to Florida. This time has gi ven me a renewed appreciation for the love a mother can give. I also need to thank my moth er for all the time she spent in raising me. She, above all others, was the most instrumental in my development and the way I think on a daily basis. I want to thank my brothe r, Brenton, and sister, Loren, for being my family in the most loving sense of the word. I cannot imagine my life without them in it. I thank my stepparents, Beverly Oakley and Thomas Qualey, for thei r love and support, and for treating me like a son. I likewise tha nk my stepsister, Jackie, for treating me like a brother, and for playing checker s every time I had a chance to visit. I would also like to
PAGE 5
v thank my grandparents, Wilmot and Ann Oakl ey and Cathy Garrett, as well as my late grandfather, Roger Garrett, fo r the encouragement and love they have given me over the years. Additionally, several aunts, uncles, a nd cousins deserve my thanks for their advice and support. I would like to acknowledge the people w ho helped mold my career choice of chemistry, and then later the choice to sp ecialize in polymer chemistry. I was first exposed to chemistry in my sophomore year in high school by Mrs. Merle Cohen, who I thank for being a wonderful teacher. The fo llowing year, I took AP chemistry from Mr. Bruce Dinsmore. By the end of his class, I had decided that I wanted a career in chemistry. I thank him for his logical teaching style, for preparing the class well for the AP exam, and for telling us more than we needed to know, as I was more than ready to take on sophomore organic chemistry as a freshman in college. At Rensselaer Polytechnic Institute (RP I), I met Dr. Brian Benicewicz, whom I thank for being the person most responsible fo r focusing my interest in chemistry towards polymers. He was also the person who first su ggested the University of Florida to me as a place for graduate study. His enthusiasm for polymer chemistry truly inspired me, and I thank him for being a very approachable res earch advisor and a friend. Other professors at RPI that I would like to thank for their he lp and guidance include Dr. R. A. Bailey, Dr. Tom Apple, Dr. Charles Gillies, Dr. Sonja Kr ause, Dr. Leonard Interrante, Dr. Wilfredo Colon, and Dr. Kim Fortun. During my junior year at Rensselaer, I was fortunate to be able to spend eight months on a co-op with the Bayer Corporation. I would like to thank Dr. Ray Burk, Dr. Wan Wu, David Jenney, Dr Inny Kim, Dr. Laura Spangler, and Tim Rappl for a wonderful experience, which pi qued my appetite for the polymer industry.
PAGE 6
vi Ray and Laura deserve extra thanks for thei r encouragement in applying to graduate schools. At the University of Florida, I would like to thank all my coworkers on the George and Josephine Butler Polymer Floor. Special th anks needs to be given to Stephen “Ed” Lehman, Jr.; Barry C. Thompson; and Ben R eeves. Ed was a fellow RPI alumni and my mentor in the laboratory, and I thank him fo r all the advice he has given me over the years. Barry, my best friend and best man, has also been an invaluable resource of knowledge in the lab, and has been a huge source of encouragement at every milestone in achieving my degree. Barry, Ben, and I starte d graduate school together and supported each other through classes, T-Hud, cumes, or al exams, and dissert ation research, and I thank them both for their help and support in al l these obstacles. Other floor members to whom I owe extra thanks for their help along the way are Jason Smith, Carl Gaupp, Jim Pawlow, John Schwendeman, Pat O’Donnell Tim Hopkins, John Sworen, Shane Waybright, Avni Argun, Stephen Carino, Andr ew Skolnik, Travis Baughman, Florence Courchay, Piotr Matloka, Genay Jones, Christ ophe Grenier, Aubry Dyer, Emine Boz, and Bob Brookins. Enough thanks cannot be expressed to Lorraine Williams and Sara Klosser in the polymer office. Without them, the little thin gs that make everything run smoothly, but no one wants to think about, would never get done. I also extend my thanks to Lori Clark and Dr. Jim Deyrup for their help in the gradua te office. I am extremely grateful to Dr. George B. Butler and his wife Josephine fo r establishing the Butler Polymer Laboratory. It has truly been a unique and wonderful place to work and study.
PAGE 7
vii Many thanks are extended to my committ ee members, Dr. John R. Reynolds, Dr. J. Eric Enholm, Dr. Danial R. Talham, and Dr. Anthony B. Brennan. Their efforts in reading and discussing the materi al presented in this disserta tion have been instrumental in making it a stronger document. Finally, I must express my sincerest appr eciation to my research director and supervisory committee chair, Dr. Kenneth B. Wagener. His door was always open and whenever I talked with him, I always felt en couraged, with my motiv ation to work hard revitalized. “Don’t give up!” he would say, when things se emed the most bleak. Never once have I seen him appear angry, as he al ways deals with situations in a calm and collected manner. Dr. Wagener has given me a wealth of advice and guidance for things both inside and outside the laboratory. I will forever be thankful for his presence in my life.
PAGE 8
viii TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES.............................................................................................................xi LIST OF FIGURES..........................................................................................................xii ABSTRACT....................................................................................................................... xv CHAPTER 1 INTRODUCTION TO SOLID-ST ATE POLYMERIZATION AND OLEFIN METATHESIS..............................................................................................1 1.1 Solid-State Reactions vs. Liquid-State Reactions..................................................1 1.2 Crystal-to-Crystal Solid-State Polymerization.......................................................4 1.3 Solid-State Polycondensation of Semicrystalline Polymers.................................12 1.3.1 Monomer-to-Polymer Condensation SSP..................................................13 1.3.2 Prepolymer-to-Polymer Condensation SSP...............................................15 1.4 Olefin Metathesis..................................................................................................23 1.4.1 Historical Overview of the Mechanism......................................................26 1.4.2 Development of Well-Defined Catalysts....................................................31 1.5 Opportunities of Olefin Meta thesis in the Solid State..........................................44 2 EXPERIMENTAL METHODS.................................................................................47 2.1 The Solid-State Apparatus....................................................................................47 2.2 Making a Solid Mixture of Reactants...................................................................48 2.2.1 Melt Polymerization...................................................................................49 2.2.2 Freeze-Drying.............................................................................................50 2.2.3 Grinding Together Diene and Catalyst.......................................................51 2.2.4 Regrinding Clean Polymer with Fresh Catalyst.........................................52 2.3 Sampling of the Solid Mixtures............................................................................52 2.4 Characterization Techniques Used on the Samples..............................................55 2.4.1 Gel Permeation Chromatography (GPC)....................................................55 2.4.2 Nuclear Magnetic Resonance (NMR) Spectroscopy..................................56
PAGE 9
ix 3 SOLID-STATE ADMET OF SE MICRYSTALLINE POLYMERS.........................58 3.1 Introduction...........................................................................................................58 3.2 Experimental.........................................................................................................62 3.2.1 General.......................................................................................................62 3.2.2 Monomers Syntheses..................................................................................63 3.2.2.1 Synthesis of tricos a-1,22-dien-12-ol (13).........................................63 3.2.2.2 Synthesis of tricosa-1,22-dien-12-one (15)......................................64 3.2.2.3 Synthesis of heneicos a-1,20-dien-11-one (16).................................65 3.2.2.4 Synthesis of the polystyre ne-graft macromonomer (19)..................66 3.2.3 Polymer Syntheses......................................................................................67 3.2.4 Other Molecules Synthesized.....................................................................67 3.2.4.1 Synthesis of undec-10-en-1-ol (21)..................................................68 3.2.4.2 Synthesis of 11-bromo-undec-1-ene (22).........................................68 3.2.4.3 Synthesis of 11-chloro-undec-1-ene (23).........................................69 3.3 Results and Discussion.........................................................................................69 3.3.1 Prepolymer-to-Polymer Experiments.........................................................69 3.3.1.1 Solid-state polymeriz ation of 1,9-decadiene....................................69 3.3.1.2 Alcohol and ketone functionalized monomers.................................78 3.3.1.3 Polystryene-graft macromonomers..................................................84 3.3.2 Monomer-to-Polymer Experiments............................................................85 3.3.3 Functional-Group Diffusion.......................................................................89 3.3.4 Thermal Analysis........................................................................................91 3.4 Conclusions...........................................................................................................97 4 SOLID-STATE ADMET OF RIGID-ROD POLYMERS.......................................100 4.1 Introduction.........................................................................................................100 4.2 Experimental.......................................................................................................104 4.2.1 General.....................................................................................................104 4.2.2 Monomer Syntheses.................................................................................104 4.2.2.1 Synthesis of 2,5-dibromohydroquinone (25).................................104 4.2.2.2 Synthesis of 1,4-dibromo-2,5-dipropoxybenzene (26)..................105 4.2.2.3 Synthesis of 2,5-dipropoxyt erephthalaldehyde (27)......................106 4.2.2.4 Synthesis of 1,4-dipropoxy-2,5-divinylbenzene (28).....................106 4.2.2.5 Synthesis of 1,4-dibromo-2,5dioctadecyloxybenzene (29)...........107 4.2.2.6 Synthesis of 1,4-dioctadecy loxy-2,5-divinylbenzene (30).............108 4.2.3 Solid-State Polymerization Conditions....................................................109 4.3 Results and Discussion.......................................................................................110 4.4 Conclusions.........................................................................................................113 5 SOLID-STATE RING-CLOSING METATHESIS.................................................114 5.1 Introduction.........................................................................................................114 5.2 Experimental.......................................................................................................117 5.2.1 General.....................................................................................................117 5.2.2 Synthesis of Dienes for Solid-State RCM................................................117
PAGE 10
x 5.2.2.1 Synthesis of N N -diallyl-3,5-dinitrobenzamide (31)......................117 5.2.2.2 Synthesis of 1-allyl-3-butenyl octadecanoate (32).........................118 5.2.2.3 Synthesis of diallyl malonic acid (33)............................................119 5.2.2.4 Synthesis of the dipotassium sa lt of diallyl malonic acid (34).......119 5.2.3 Solid-State RCM Reactions Conditions...................................................120 5.2.4 Synthesis of Ring-Closed Prod uct (2,5-dihydropyrrol-1-yl)-(3,5dinitrophenyl)-methanone (35)................................................................121 5.3 Results and Discussion.......................................................................................121 5.4 Conclusions.........................................................................................................127 LIST OF REFERENCES.................................................................................................129 BIOGRAPHICAL SKETCH...........................................................................................140
PAGE 11
xi LIST OF TABLES Table page 1-1 Functional group tolerance of early and late transition metal olefin metathesis catalysts..................................................................................................34 1-2 Comparison of condensate boiling points................................................................46 3-1 Results of solid-state polymerizations of monomer 11 with periodic sampling...................................................................................................................70 3-2 Useful molecular weights of typical step-growth polymers.....................................71 3-3 Results of solid-state polymeri zations using SchrockÂ’s catalyst 2 and monomer 11 ..............................................................................................................72 3-4 Catalyst loading comparison using monomer 11 and catalyst 6 ..............................73 3-5 Experiments conducted at 60 C using monomer 11 ...............................................75 3-6 Comparison of argon and vac uum experiments using monomer 11 and catalyst 6 ...................................................................................................................77 3-7 Examples of solid-state polymerizations of alcohol monomer 13 ...........................80 3-8 Examples of solid-state poly merizations of ketone monomer 16 ............................81 3-9 Precipitation effects on the solid-sta te polymerizations with catalyst 5 and ketone 16 ...........................................................................................................83 3-10 Results from the monomer-topolymer freeze-drying experiment using ketone 15 .........................................................................................................86 3-11 Results from the monomer-to-pol ymer experiments using ketone 16 and catalyst 6 (100:1 monomer to catalyst ratio).....................................................87 3-12 Results of SSPs starti ng from a prepolymer of 18 without a tail on the melting endotherm that extends in to the reaction temperature................................95 5-1 Results of solid-state RCM experiments................................................................123
PAGE 12
xii LIST OF FIGURES Figure page 1-1 Topochemical solid-state polymerization..................................................................5 1-2 The two types of topochemical SSP: (a) homogeneous growth starting at randomly distributed point in the crystal; (b) heterogeneous growth by nucleation..............................................................................................................6 1-3 Photopolymerization of diacetylenes.........................................................................8 1-4 Projection of the monomer and pol ymer crystal structures of the bis( p -toluenesulfonate) diyne (TS).............................................................................9 1-5 Changes in unit cell parameters for PTS as the polymerization proceeds.................9 1-6 Photopolymerization 2,5 -distyrylpyrazine (DSP)....................................................10 1-7 Conversion of (a) DSP into (b) dimer and (c) polymer............................................11 1-8 Solid-state monomer-to-polymer polycondensation of 6-aminohexanoic acid and 11-aminoundecanoic acid to Nylon 6 and Nylon 11, respectively............14 1-9 Formation of a prepolymer of PET followed by an increase in the degree of polymerization while in the solid state................................................................16 1-10 Reactive functional groups in the amorphous region that are anchored in the crystal.............................................................................................................18 1-11 Typical plot of mol ecular weight versus time for a polycondensation SSP.............21 1-12 Transesterification intercha nge reaction taking place along the polymer backbone....................................................................................................22 1-13 Modes of olefin metathesis in both organic and polymer chemistry.......................25 1-14 Early-proposed quasicyclobutane m echanism for olefin metathesis, which was later found to be incorrect......................................................................27 1-15 Tetramethylene metallocycle mechan ism for olefin metathesis, which was later found to be incorrect.................................................................................28
PAGE 13
xiii 1-16 Metal-carbene mechanism for olefin metathesis, which is now well established as the correct mechanism.......................................................................30 1-17 Catalytic cycle for olefin me tathesis in an example of RCM..................................31 1-18 Common version of SchrockÂ’s tungsten ( 1 ) and molybdenum ( 2 ) based olefin metathesis catalysts........................................................................................33 1-19 Select well-defined ruth enium metathesis catalyst..................................................35 1-20 Associative and dissocia tive pathways of first( 5 ) and second( 6 ) generation Grubbs catalyst.......................................................................................39 1-21 HoveydaÂ’s catalyst undergoi ng RCM, whereby the original oxygen-coordinated catalyst is reformed.................................................................40 1-22 Bimolecular decomposition of catalyst 6 to form diruthenium complex 10 which is likely responsible for isomerization during olefin metathesis.......................................................................................................43 1-23 Mechanism proposed by Grubbs for the decomposition of catalyst 6 .....................44 1-24 The ADMET mechanism whereby the equilibrium is shifted toward the polymer by the release of ethylene.....................................................................45 2-1 Drawing of a typical apparatus for solid-state reactions..........................................48 2-2 Freeze-drying apparatus with liquid nitrogen trap...................................................51 2-3 Reaction of rutheniu m catalysts with EVE..............................................................53 2-4 Free radical scavenging with BHT...........................................................................53 2-5 Relative NMR integrations of inte rnal and external olefins for NMR end-group analysis....................................................................................................57 2-6 Olefin region in the NMR for typical ADMET polymers with long methylene spacers....................................................................................................57 3-1 Monomer used in solid-state ADMET polymerization............................................61 3-2 Gel Permeation Chromatography traces of the SSP of 1,9-decadiene with first-generation Grubbs at 45 C sampled periodically....................................71 3-3 Prepolymerization in the melt of 1,9-decadiene, followed by polymerization in the solid state using second-gene ration Grubbs ruthenium catalyst ( 6 )...............77 3-4 Synthetic route to alcohol 13 and ketone 15 ............................................................78
PAGE 14
xiv 3-5 Mechanism of synthesis of ketone 16 ......................................................................79 3-6 Atom transfer radical poly merization (ATRP) of initiator 24 to macromonomer 19 with subsequent ADMET to polymer 20 ..................................85 3-7 The ADMET mechanism at work in the amorphous region of a semicrystalline solid.................................................................................................90 3-8 Typical DSC showing the envi ronment at the start of a 1,9-decadiene ( 11 ) experiment.................................................................................92 3-9 A DSC of a precipitated prepolymer from monomer 16 ..........................................92 3-10 Both the DSC scans of the starting and 7-d samples from the 1,9-decadiene experiment in Figure 3-3...................................................................93 3-11 A DCS trace of ketone polymer 18 that was synthesized in solution, precipitated, and washed with boiling methanol to eliminate the tail on the melt endotherm..............................................................................................94 3-12 Typical GPC curve of polymer from monomer 11 and catalyst 6 ...........................97 4-1 Indirect synthesis of PPV vi a a soluble polymer precursors established by Wessling.........................................................................................102 4-2 Synthetic route to 1,4-d ipropoxy-2,5-divinylbenzene ( 28 )....................................110 4-3 Synthetic route to 1,4-dioct adecyloxy-2,5-divinylbenzene ( 30 )............................111 4-4 Prepolymerization in the melt of 28 followed by polymerization in the solid state using second-generation Grubbs ruthenium catalyst ( 6 ).................112 5-1 Comparison of the ADMET and RCM mechanisms.............................................115 5-2 Common RCM dienes found in the l iterature for testing features of RCM (E = CO2Et).............................................................................................116 5-3 Diene synthesized for solid-state RCM experiments.............................................122 5-4 Ring-closing metath esis of amide diene 31 to compound 35 using catalyst 5 .................................................................................................................124
PAGE 15
xv Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy SOLID-STATE OLEFIN METATHESIS By Garrett W. Oakley December 2004 Chair: Kenneth B. Wagener Major Department: Chemistry While organic reactions are most often pe rformed in the liquid or melt state, the investigation of solid-state or ganic conversions and solid-sta te polymerization (SSP) have received varying degrees of attention for several decades now. Solid-state polymerization is traditionally divided into two classes: crystal-to-crystal polymerization, well established by Wegner; and equilibri um polycondensation of semicrystalline polymers, which is widely used in industry. The principles governing these two types of polymerization are quite different, for the fo rmer is based on topological chemistry, while the latter is thought to occu r in the amorphous region of semicrystalline polymers. Olefin metathesis is an astonishing chemical phenomenon, whereby a single chemical reaction has become widely used and st udied in several, very different fields of chemistry, in both academic and industrial settin gs. The principle mission of this work is to demonstrate that a particular mode of olefin metathesis, acyclic diene metathesis (ADMET), can operate in the solid state using the same principles used for semicrystalline condensation polymers in industry. Experiments were aimed at
PAGE 16
xvi mimicking the industrial proces ses, including a prepolymer ization period in the melt state, extrusion of this prepolymer to a smalle r particle size, and flow of an inert gas (or vacuum) through the reaction vessel to aid in condensate removal. In most cases of ADMET, the condensate that is released is et hylene, which gives it a distinct advantage over polyester, polyamide, a nd polycarbonate in dustrial processes in that the condensate is extremely volatile (BP = –104 C), and therefore high temper atures are not required for condensate removal to drive the equilibrium toward high polymer. The “benchmark” hydrocarbon monomer, 1,9-de cadiene, was chosen for the initial work because it lacks any polar or protic func tionality that could react or complex with the metathesis catalysts. Steady increase in molecular weight was observed in the solid state, as low as room temperature, w ith the firstand second-generation Grubbs ruthenium catalysts. Similar trends of mo lecular weight increase were observed with monomers of varying functiona lity and architecture, which included ketone and alcohol functionalized monomers and a polystyrene gr aft copolymer. Conclu sions with potential industrial implications were drawn from therma l analysis. We believe that low molecular weight oligomer and unreacted monomer act as plasticizers for the promotion of solidstate polymerization. In experiments wher e these species were removed, no significant molecular weight increase was observed. Additionally, solid-state ADMET of rigid-ro d polymers and ring-cl osing metathesis (RCM) were explored. Dial koxy-divinylbenzene monomers we re synthesized and used in ADMET polymerizations, where slight increases of molecular weight in the solid state were observed. Solid-state RCM to low conv ersions was also demonstrated with an
PAGE 17
xvii amide diene before a subsequent, visible, he terogeneous change, which we interpreted as local melting point depression about sites of defects from the presence of RCM product.
PAGE 18
1 CHAPTER 1 INTRODUCTION TO SOLID-STATE POLYMERIZATION AND OLEFIN METATHESIS 1.1 Solid-State Reactions vs Liquid-State Reactions Most organic reactions are performed in the liquid or melt state because these states possess the relevant molecular freedoms to a llow the reaction to occur. A common view among chemists, and even x-ray crystallographers, is that a molecular crystal is made up of frozen chemical entities, which can only e xperience slight molecu lar motions. Indeed this is true for many cases; however, a vast assortment of chemical reactions do occur in the solid state and may have several advant ages over performing th e reactions in the liquid state. These reactions have received varying degrees of attention in the literature for several decades now.1,2 In some cases, a reaction may not even occur in the liquid state, and the solid state provide s the only route to the product. When considering a reaction in the solid st ate, several factors not present in the liquid state must be examined. In the cas e of a solid-state pol ymerization (SSP) for example, some of these factors are the cr ystallinity of the m onomer and polymer, the polymorphic phase(s), nucleation an d growth, sites of defects in the crystal, and particle size. The one factor, however, that plays the most dominant role in the feasibility of any solid-state reaction is the abil ity of the reacting groups to interact. Whether the solid environment is crystalline, semicrystalline, or amorphous in nature, there must be intimate contact between reacting groups in order for the reaction to occur.
PAGE 19
2 This is extremely important when working within a crystalline environment, where only very slight vibrational motions are a llowed. For the reaction to occur, the two reacting groups must be very close in pr oximity (~4 ) and must have the proper orientation in order for the reaction to occur.3 It is an extremely fortuitous when a reaction of this nature can proceed in a crysta lline state because the crystal structure of a molecule is not something that is chosen by th e scientist. Mother Nature decides the best way for a particular molecule to stack togeth er, and that structure will never change for that particular molecule under a standard set of conditions. It is possible to change the crystal structure by changing th e crystallization procedure, although this is quite rare. Usually, to change the crystalline structure, the molecular structure must change, thereby changing the molecule being dealt with. This is not always a flexibility that can be afforded. Changing the chemical structure of the molecule often changes the properties of the product as well. This may be very limiting depending on the desired application. Tailoring of a crystal structure to suit the need s of solid-state reaction is also impractical, since space-filling patterns are extremely difficult to predict. A slight change at one site on a molecule can have huge, unexpected ch anges in crystal stru cture, which could change its space-filling patter n, and even its lattice type. For most organic reactions where the st arting material is a solid, the crystal structure is just not conducive to performing the reaction in the solid state. In addition, multiple reagents often are needed for the r eaction to occur. Every chemist knows that for most reactions, simply mixing two solids toge ther will result in no reaction. Solvents or melting are needed to allow the reacting sp ecies the intimate contact they need for the
PAGE 20
3 reaction to occur by offering the much highe r degree of molecular motion allowed in a liquid state. Amorphous or partially amorphous enviro nments offer elevated mobility for functional groups to meet. Above the glass transition temperature ( Tg) of a polymer, the amorphous areas are softened from their glas sy state. Chains within the amorphous region may shift their position, increasing th e probability of reactive groups meeting, even if the chain is anchor ed in a crystalline region. There are several examples where so lid-state polymerization does occur.3,4,5 The main interest in studying SSP is to obtain polymers that are unattainable in the liquid state or polymers that have better properties th an those obtained in th e liquid state. One of the most desirable features of SSP is that some polymers have be en shown to go to a higher extent of reaction and achieve higher mo lecular weights than those obtained in the liquid state. It is well known that many polymeric prope rties can be improved with increase in molecular weight and SSP can provide a route to them. Solid-state polymerization may also provide a route to intractable polymers (i.e., polymers that are insoluble in almost every solvent and precip itate out of solution even at the dimer or trimer stages of the polymerization). Speci al polymer morphologies are possible because the reactions are occurring in the solid state. In many cases, the polymer created has a high degree of crystallinity and/ or stereoregularity that woul d be difficult or impossible to achieve by any other means. Unwanted side reactions that lead to undesirable byproducts are limited or avoided because of clos e packing in the crystal. It has been shown that polymers prepared by SSP often have less cyclic oligomer than those prepared by other methods.6 Formation of crystals with fewer defects, such as spurious
PAGE 21
4 moieties in the chains or chain branching, has also been demonstrated.7 Another reason for these observations, aside from being in the solid state, is that the reaction is done at lower temperatures than in the melt or solu tion. The high temperatures needed to keep the material in the melt or in solution are not needed, so fewer side reactions occur. In addition, SSP is commercially appealing because no solvent is required, there is practically no pollution, and it requires less energy th at it would to k eep the polymer in the melt state. Solid-state polymerization is traditionally divided into two classes:8 crystal-tocrystal polymerization,3,9 well established by Wegner; and equilibrium polycondensation of semicrystalline polymers,4,5 which is widely used in industry. The principles governing these two types of polymerization ar e quite different, for the former is based on topological chemistry, while the latter is thought to occur in the amorphous region of semicrystalline polymers. 1.2 Crystal-to-Crystal Solid-State Polymerization Crystal-to-crystal solid-state polymerization actually occurs in a crystalline state. Therefore, it relies on a suitable arrangement of the reactants in the monomer crystal to allow for the necessary contact between r eacting groups with only the very slight vibrations allowed in a solid crystal. These reactions occur by only very slight rotations of the reactants in a crystal (Figure 1-1). The ideal case is a si ngle-crystal-to-singlecrystal transformation of monomer to polymer The only example of SSP where this is actually known to occur is the polymeriza tion of certain diacetylenes in the solid state.9,10,11
PAGE 22
5 Figure 1-1. Topochemical so lid-state polymerization Most topological SSPs occur with a higher degree of disorder including examples with nucleation of the polymer from a site of disorder or diffusion of reactants to centers of reactivity. This can lead to complete destruction of the parent crystal with a high degree of irregularity in the da ughter crystal. However, the interesting examples retain some degree of order in the daughter crystal. If a polymer crystal obtained in the solid state retains a small number of crystallographi cally equivalent orie ntations of the monomer (usually in one direction determined by th e parent crystal structure) it is termed topotactic. In contrast, a reac tion is termed topochemical if the reaction is governed by the geometry and spatial symmetry of the adjacent molecules and the reaction pathway is uniquely defined by the parent crystal structure. Topochemical polymerization can be thought of as a diffusionless transformation of monomer single crystal to polymer single crystal. All topochemical reactions are topot actic, but there are examples of topotactic reactions that are not cons idered topochemical. One such example is ring-opening polymerization of oxacyclic compounds, such as trioxane and tetraoxane, which occurs with diffusion of monomer. These types of ex amples that have a la rge degree of disorder
PAGE 23
6 in the parent crystal will not be discussed further, but have been reviewed in the literature. 3,12 Topochemical reactions can be either hom ogeneous or heterogeneous. In a homogeneous polymerization, chain growth starts at randomly distribu ted points throughout the crystal lattice, and then forms extended ch ains of polymer in a solid solution (Figure 1-2a). Wegner,9 a pioneer in the field of the only known homogeneous SSP, describes it as “a solid solution of individua l extended chains of varying length all stretched out along a unique crystallographic direction inside the matrix of the as yet unreacted monomer.” Heterogeneous solid-state polymerizations (Figure 1-2b) are the more common type. Figure 1-2. The two types of topochemical SSP: (a) homogeneous growth starting at randomly distributed point in the cr ystal; (b) heterogeneous growth by nucleation For the heterogeneous type, the reaction preferentially starts at defect sites in the parent crystal, and proceeds with nucleati on in the daughter phase. This process of nucleation and phase growth results in fragment ation of the parent crystal, with a higher probability of nucleation at the surface of an already existing nucleus. The formation of a nucleus of an amorphous polymer inside the monomer crystal undoubtedly will cause stresses within the crystal. These stresses are relieved by the form ation of dislocations around the original nucleus, which produces a new generation of reaction sites for
PAGE 24
7 nucleation.7 This in turn will produce a multitude of small crystallites in a solid, as opposed to a large single crystal. It happens that the best examples fo r both homogeneous and heterogeneous SSP are polymerizations through unsaturated bonds. The homogeneous example is that of the diacetylenes,9,10,11 which polymerize via 1,4-addition of conjugated triple bonds; and the heterogeneous example is that of diolefins,13,14 which undergo [2+2] cycloaddition to form cyclobutane rings in the backbone of th e polymer. For a truly topochemical crystalto-crystal transformation, the unsaturated bonds in both cases must be mutually oriented in the crystal. In this mu tual orientation of the molecu les, the unsaturated bonds of adjacent molecules are close to parallel, preferably coplanar to each other, and not exceeding a distance of 4 between them.3 The polymerization of conjugated diacetylenes in the solid state occurs thermally or with high-energy photoradiation. The polym erization occurs by a 1,4-addition with diradical and dicarbene intermediates to gi ve rise to a fully conjugated main chain (Figure 1-3), where the polymer produced ha s two resonance structures. One structure has three cumulated double bonds per repeat un it, while the other has alternating double and triple bonds in conjugation. The di acetylene rods undergo slight rotational movement with subsequent addition to the nearest neighbor during the polymerization. X-Ray data has shown that in a crystal of pol ymeric diacetylene, the substituents adopt an alltrans configuration.15 The crystal packing of the monomer is determined by the kind of substituents on the diacetylene. These polymerizations can only proceed in the solid state if the substituent groups interact strongly with intermolecular hydrogen-bond formation or
PAGE 25
8 dipolar forces. The mutual position of the triple bonds in the cr ystal structure, and therefore the reactivity of th e monomer, is controlled by how closely the substituents allow the molecules to pack together. If the intermolecular forces are not strong enough to create a closely packed structure, then the polymerization will not proceed. A typical example of a diyne that under goes a successful SSP is the bisp -toluenesulfonate of 2,4-hexadiyne-1,6-diol, abbreviated TS.16 The polymer product is thus abbreviated PTS. R R R R R R R R R R R R R R R R R R h Figure 1-3. Photopolymeriz ation of diacetylenes Figure 1-4 shows a comparison of the atomic positions of the monomer and polymer of PTS. Note how there is only a ve ry slight change in the crystal structure during the course of the reacti on. This allows the polymeri zation to proceed without any appreciable change in the original shape an d size of the monomer cr ystal. Figure 1-5 illustrates the very small changes in unit cell parameters during the course of the reaction. Molecular weights of PTS reach orders of ~106 g/mol. At 100 % conversion, the distance between chain ends is very small (less than 3 ); therefore, the polymer chains can be considered as continuous lines extending from one end of a cr ystal to the other. Consequently, the product of a polymerization of a diacetylene occurring in this manner is considered a polymeric single crystal. This is the only polymer chemically occurring in the form of a macroscopic single crys tal; hence it is the only known homogenous SSP.
PAGE 26
9 TSSOCH2O O CH2OS O O Figure 1-4. Projection of the monomer a nd polymer crystal structures of the bis( p toluenesulfonate) diyne (TS) Figure 1-5. Changes in unit cell parameters for PTS as the polymerization proceeds A typical example of a heterogeneous SSP is the [2+2] photopolymerization of diolefines.13,14 In contrast with the topochemical polymerization of di acetylenes, this reaction proceeds by a step-growth mechanism. Upon UV radiation of wavelengths of 450-350 nm, double bonds on adjacent molecules undergo a [2+2] photocycloaddition to
PAGE 27
10 form cyclobutane rings within the polymer backbone. The monomer does not necessarily move much during the polymer ization. However, this slight movement causes the reaction to proceed by phase separation with nu clei growth that can advance in different crystallographic directions at different speeds. This is opposed to chain growth in one crystallographic direction, as in the case of a homogeneous diac etylene solid-state polymerization. Although the result is destruc tion of the parent crystal, separate welldefined phases can be observed in the crysta l even in the early stages of the reaction.17 The first example of this four-center type photopolymerization of a crystal to be studied was the polymerization of 2,5-distyrylpyrazine (DSP)18 (Figure 1-6). The result of the polymerization is a highly crystallin e linear polymer produced in quantitative yield. N N N N n h DSP poly-DSP [2+2] Figure 1-6. Photopolymerizati on 2,5-distyrylpyrazine (DSP) The change in crystal structure as the polymerization proceeds is illustrated by Figure 1-7. The distance between the reacti ng olefins in the parent crystal has been found to be 3.939 ,19 which is within the 4 criteria discussed earlier for a truly topochemical reaction to occur.
PAGE 28
11 Figure 1-7. Conversion of (a) DSP into (b) dimer and (c) polymer In some other cases of diolefins that could potentially undergo the [2+2] cycloaddition, the propagation is hindered by an in crease in the distan ce between reaction sites, even after dimer is formed, due to th e molecular displacement of the reaction. The double bonds become too far apart for th e reaction to propagate further. The symmetry of the molecular arrangement is preserved during the course of the polymerization of DSP, as indicated by the coincidence of the space groups ( Pbca ) of the monomer and the polymer crystals. The reactio n is therefore termed a lattice-duplicating polymerization. The only significant change in the crystal structure is a 1.8 % contraction in the c direction from monomer to polymer.12 Of the several other types of heterogeneous SSP, a few are worth mentioning. Polymerization of 1,3-dienes via a 1,4-additio n in the solid state has recently been reviewed.20 Some vinyl monomers can undergo a radical chain mechanism in the solid state with a high degree of disorder.7 Lastly, there are even some examples of SSP of the
PAGE 29
12 polycondensation type that transform from crysta l-to-crystal in the solid state. One such example is the formation of polyesters th rough thermally induced polymerization of halogenoacetates.5 1.3 Solid-State Polycondensation of Semicrystalline Polymers Solid-state polycondensation of equilibrium step-growth polymers was first observed more than 40 years ago.21,22,23,24 This type of SSP is di stinctly different from most topochemical SSPs. One clear difference is that a small molecule is released from the reaction for every propagation step that occurs. Since many polycondensation reactions are reversible, this condensate must be removed in order to drive the equilibrium toward the polymer and achieve hi gh molecular weight. The diffusion of this side product to the surface of the particle can become the ra te-limiting factor if there is not sufficient surface area to allow the condens ate to escape the solid. The condensate is most frequently removed by static or dynamic vacuum or by a stream of an inert gas. Recently, it has been show n that supercritical CO2 may also be used as a sweep fluid.25,26,27,28 Another distinct difference is that these are usually semicrystalline polymers, where the polymerization is conducted abov e the glass transition temperature ( Tg), but below the melting point ( Tm) of the polymer. This greatly enhances molecular mobility within the amorphous region of the polym er, which facilitates end-gr oup encounters and aids in condensate diffusion. Since the reaction is co nducted below the semicrystalline melting point, the crystalline regions remain intact, a nd the particle maintains its shape during the polymerization. There are two distinct star ting points for polycondensatio n reactions in the solid state. The atypical case starts with a crys talline monomer, and goes all the way to the
PAGE 30
13 polymer in the solid state. This is a rare occurrence, but it does potentially have great scientific and commercial interest. More frequently however, the starting material has already been polymerized to some extent in the melt or solution state, and is further polymerized as a solid, leading to higher molecular weight. To term this type of polymerization as a “solid-state polymerization” causes discontent among some scientis ts. For one, the reaction is occurring above the glass transition temperature of semicrystalline polymers, so there is a higher amount of mobility in the amorphous region than in a crystalline solid, which many scientists consider to be the “solid-state.” Additiona lly, after the prepolymerization in the melt, there may only be few couplings per chain ne eded in the “solid-state” portion of the process to achieve the desired, higher mol ecular weight. The word, “polymerization,” by definition, means that many couplings are taki ng place in the process. However, to not call this a “solid-state polymerization” will co ntradict more than 40 years of literature and industrial terminology. A more suitable definition for this t ype of SSP is a step-growth polymerization of a semicrysta lline polymer, where the crys talline region maintains the shape of particle while the polymer is growing. 1.3.1 Monomer-to-Polymer Condensation SSP Before a thorough considerati on of the solid-state polymer ization of prepolymers, further mention of monomer-t o-polymer polycondensation is worthwhile. These examples span the two classes of SSP: crysta l-to-crystal and polyc ondensation. Like any crystal-to-crystal SSP, reactiv ity and the kind of crystal line order in the product are determined to a considerable degree by the characteristics of the parent crystal. Again, the parent crystal for monomer must have an acceptable arrangement of molecules, in which the reacting functional groups are in close enough proximity and have the proper
PAGE 31
14 orientation to allow the reaction to occur in the solid state. It is remarkable, however, that this occurs despite the need to provi de for the diffusion of low molecular weight molecules to the surface of the particle. Re ductions in volume a nd contractions of a crystal along the direction of the chain have been observed during some SSPs from the elimination of the condensate. One such example is the polymerization of 6-aminohexanoic acid to Nylon 6 (Figure 1-8).29 It was shown to have an 18 % reduction in volume and a 17 % contraction in the direction of the polymer chain. Polymerization of 11-aminoundecanoic acid to Nylon 11 (Figure 18) was found to have a similar effect with 16 % reduction in volume and 11 % contraction.4 O OH H2N O H N n xx x = 1 6-Aminohexanoic acid 5 11-Aminoundecanoic acid x = 1 Nylon 6 5 Nylon 11 160-170 oC vacuum +n H2O Figure 1-8. Solid-state m onomer-to-polymer polycondensat ion of 6-aminohexanoic acid and 11-aminoundecanoic acid to Nylon 6 and Nylon 11, respectively Some monomers have been shown to polymer ize to some extent in the solid state, but not to the same degree of polymerizatio n that can be achieved in the melt. An example of this is polymerization of bis( 2-hydroxyethyl terephth alate) to produce poly(ethylene terephthalate) (PET; Figure 1-9). An example of this is polymerization of bis(2-hydroxyethyl terephthala te) to produce poly(ethylene te rephthalate) (PET; Figure 1-9). Poly(ethylene terepht halate) is a typical exampl e of polycondensation SSP by which a prepolymer of low molecular wei ght can achieve a hi gher degree of polymerization in the solid state, 30 as discussed next. However, if starting a solid-state polymerization from a crystal of the monomer, it is very slow indeed, and it is virtually impossible to achieve high molecular weight starting from monomer. Bamford and
PAGE 32
15 Wayne30 reported dimer as the primary product, both with and without catalyst, based on melting-point correlations. 1.3.2 Prepolymer-to-Polymer Condensation SSP Solid-state polycondensation of semicrystall ine polymers is widely practiced in the polyester, polyamide, and polycarbonate industr ies. Typically, the polymerization begins in the liquid (melt) state at hi gh temperatures (usually above 250 C) to achieve a moderate molecular weight, and then the polymer ization is continued in the solid state to achieve higher molecular weight. High molecular weights, often unattainable in the molten state, can be obtained by SSP of pr epolymers, leading to desirable physical properties. Nylon 6,6 has been shown to achieve molecular weights of 280,000 g/mol from starting material of 40,000 g/mol; a 7-fold increase in molecular weight while in the solid-state.31 For the polyester, poly(ethyl ene terephthalate) (PET), th e monomer is synthesized by reaction of the terephthalic acid or its dime thyl ester with ethylene glycol to generate the bis(2-hydroxyethyl) ester. This is done to alleviate th e need for a stoichiometric balance of two monomers, as is the case in mo st step polymerizations. The chemistry is a simple transesterification that is reversible if the by-product, ethy lene glycol, is not removed. Figure 1-9 shows the process by which this polymer is synthe sized. First, the monomer is polymerized in the melt. The molecular weight achieved during this first process is approximately 15,000 g/mol, which is high enough for use in most applications. This polymer is then cooled to a temperature below its semicrystalline melting point, and extruded into small particle s to increase its surface area. The polymer
PAGE 33
16 chemistry is then continued in the solid stat e above the glass transition temperature, but below the melting point of the semicrystalline polymer (usually about 150 C) to achieve higher molecular weight. A molecular wei ght of approximately 25,000 g/mol can be achieved for PET in the solid state, which can be used in more demanding applications. HOO O O O ~280 oC (melt) OH HOOH + OO O O x Tg < Temp < TmHOOH +y >> x(solid state) Prepolymer Nitrogen Flow OO O O x OO O O y Vacuum Figure 1-9. Formation of a pr epolymer of PET followed by an increase in the degree of polymerization while in the solid state Completing the polymerization at a lower te mperature has distinct advantages over performing the polymerization completely in th e melt state. It is economically practical because less energy is required to heat th e reaction mixture, and because polymers of higher purity are produced due to fewer side reactions occurring, a nd this can, in turn, lead to higher molecular weights and pr operties unattainable in the bulk state. As in many engineering plastics, molecu lar weight plays an important role in achieving the desired properties of the polymer. Certain processes, such as blow-molding and extrusion, require that the molten pol ymer has a high enough melt viscosity to prevent collapse in the soft preformed state. This requires high molecular weights, which can be obtained only with great difficulty in conventional melt processes. To achieve this in the melt, extremely high temperatures are ne eded to lower the viscosity in order for the
PAGE 34
17 reaction to occur, and therefore side reactions occur that counteract the desired increase in molecular weight. In addition, polymer me lts with high melt viscos ities are difficult to handle and hard to remove from bulk polymerization equipment. These disadvantages can be simply avoided if SSP is an option. Solid-state polymerization of semicrystalline prepolymers is a complex process. The prepolymer is extruded into small chips or a powder. Tumble dryers or moving bed reactors are usually employed to keep the sm all particles from sticking together. The polymerization is continued at a reaction te mperature well above the glass transition temperature of the polymer, but below its melting point ( Tg < Temp < Tm). A typical SSP takes place 5-40 C lower than the melting point of the polymer.4 The small particles are then subjected to a stream of inert gas or vacuum to remove the condensate. Work has been done comparing the two;32 however, the choice is usually based on convenience. Removing the condensate at the surface of the particle drives the equilibrium toward a higher molecular weight polymer, improving the physical properties. Usually, the shape of the material is maintained during a solid-state polymerization, and then the product polymer can be extruded in to its desired shape. If the polymer is molded into its final form after the polymer ization in the solid state, it is termed a postextrusion SSP process. This is the on ly option unless the desired sculpt of the product has at least one dimension small enough to allow the condensate to escape. Then a preextrusion SSP process is possible, in wh ich the prepolymer is fashioned into its desired shape first, and then further polymerized in the solid state while maintaining the same shape. Such can be the case in thin films and fibers.
PAGE 35
18 Several factors play a role in the ove rall reaction kinetics of a solid-state polycondensation. These include the crystallinity of the polymer, par ticle size, initial molecular weight, number and type of chai n end groups, reaction temperature, time of polymerization, diffusion of condensate to the surface of the particle and technique used to remove the condensate at the surface. Most reports in the literature only treat a few of these at a time, mainly treating the factors th at affect the overall rate of the process. It is generally accepted, but not proven, that the polymerization occurs in the amorphous region. The rate of reaction is therefore restricted by the molecular mobility of the reactive end groups, which are anc hored down in the crystalline phase shown by Figure 1-10. Figure 1-10. Reactive functiona l groups in the amorphous regi on that are anchored in the crystal In the kinetic treatment of data, it is sometimes assume d that all the end groups are in the amorphous region to explain an increas e of crystallinity with increased molecular weight. This is supported by a study on the degree of crystallinity of polyamides at a constant pressure.33 If this assumption is correct, th e concentration of end groups in the
PAGE 36
19 amorphous region would be higher than if they were dist ributed throughout the entire volume of the solid, and the reaction would be shifted towards the polymer in order to reestablish equilibrium. The result is an increase in molecular weight. The degree of crystallinity ma y also have an effect on the diffusion coefficients of the condensate.24,34 The higher the degree of crystallinity, the harder it is for the condensate to diffuse through the solid. This is coupled with the fact that the degree of crystallinity of some polymers increases as the solid-state polymerization proceeds.34 Newly formed chain extensions at the surface of the crystalline phase can be taken into the crystal, increasing the pe rcent crystallinity. This phenomenon is called SSP-induced crystallization. The surface of larger particles will become more crystalline than the inner portions of the particle since the condensate can more eas ily escape. This, in turn, makes it even more difficult for the condensat e to diffuse to the surface and escape from the inner portions of large particles. The particle size is the most unique factor to play a role in polycondensation SSP. In general, while the overall rate incr eases with a decrease in particle size,30 the kinetics can be much more complicated. With extrude d geometries that allow for ease of escape of the condensate in at least one dimension, diffusion of this side product to the surface becomes a negligible part of the picture. The condensate does not have an opportunity to undergo depolymerization if it does not have the opportunity to collide with a polymer molecule. At larger sizes, ot her factors, such as temperature, play a larger role. Poly(ethylene terephthalate), for example, at sizes greate r than 100 mesh is diffusion controlled at 210 C. However, at 160 C, the chemical reaction is slower, and the chemical kinetics becomes the rate-limiting factor This suggests that at temperatures in
PAGE 37
20 the range of 160-210 C, the kinetics are c ontrolled by both diffusion and chemical kinetics.24 At particle sizes where the reaction is di ffusion controlled, sections of a polymer chip have been shown to have decreasing intrin sic viscosity as penetr ation into the center of the polymer chip increases.30 This is evidence that the reaction is occurring primarily at the surface of the particle since the higher molecular weig ht portions are more viscous. Broadening of the overall molecular weight di stributions have also been observed with increased particle size.35 Larger particles will have a br oader molecular weight gradient in the polymer chip, causing the overall molecula r weight distribution of the sample to be broader. Choice of temperature is one of the most im portant factors because it affects almost all other aspects of th e reaction. Too high a temperature may lead to partial melting with concurrent sticking of the polymer chips or a change in the nature of the process, while too low temperatures may prevent chain grow th or lead to long reaction times. The temperature for most SSPs is very high due to the high activation energy of the overall process. Typically, the temperature of the SSP is 5-40 C lower than the melting point of the polymer to overcome this activation energy.4 The lowest temperatures where SSP takes place in a detectable manner are a pproximately 150-170 C for polyesters and 90-120 C for polyamides. The time of the reaction is another important factor. A typical plot for a SSP of the molecular weight vs. time is shown in Fi gure 1-11. The molecular weight reaches a maximum and eventually drops down due to the occurrence of degradation reactions.
PAGE 38
21 Optimization of reaction time must be done to achieve the most desirable molecular weights in a SSP. Figure 1-11. Typical plot of molecular weight versus ti me for a polycondensation SSP As in any reaction, much is dependent on the nature of the reagents used. The initial molecular weight and the number a nd type of end groups of the prepolymer determine what kind of molecular species will be involved in the reaction. The mobility of the catalyst employed can also be a factor Titanium catalysts have been used in polyester synthesis as a coordination catalyst.36 It might be that the larger the amount of catalyst in the mixture, the less diffusion of th e catalyst is required for it to be effective. Much of the discussion regarding solid-s tate polycondensation thus far has been centered on the diffusion of some species throu gh a solid. In addition to the diffusion of small molecules to the surface, the migration of end-group functionality also needs to be considered. Most models that consider the migration of end groups rely on the diffusion of segments of polymer chains within the solid in order for end groups to meet and react. However, rapid buildup of molecular wei ght cannot be rationalized by diffusion of molecules or their segments alone, since migration of end groups would be required for frequent encounters of reactive groups.31 Thermal motions are severely restricted in the
PAGE 39
22 solid state, and segmental diffusion is consid erably restricted as compared to the melt state. In addition, the motions of most chain segments in semicrystalline polymers are restrained by the fact that they are anchored in the crystal. A more recent model suggests that end group migration can be explained by interchang e reactions. Although not recognized by most as the dominant mechan ism, this model has received some acknowledgment in recent years. 31,37 In a polycondensation reacti on, if the end group reacts w ith another end group, then a small molecule is released. However, an end group also has the ability to react with any functional groups along the chain backbone in an interchange re action, such as the transesterification shown in Figu re 1-12. The result is just an exchange of functionality and migration of a chain end with no affect on the number average molecular weight. However, these exchanges of functionality facilitate the meeting of two end groups, which can, upon reaction, release a small molecu le and drive the reaction toward a higher molecular weight. O O R O O R OH O O O O O O O O R O O O O O O R R OH R OH OR Figure 1-12. Transesterifica tion interchange reaction taking place along the polymer backbone There is significant indirect evidence th at these exchange reactions occur during a SSP. Lenz and coworkers38,39 studied reorganizatio n of copolymers, mostly copolyesters,
PAGE 40
23 which showed an increase in the extent of blockiness of the c opolymer upon heating in the solid state. This requires that intercha nge reactions are taking place. A semicrystalline copolymer can undergo a reaction-faci litated crystallizat ion and develop an increasingly blocky structure if its repeat units can undergo interchange reactions. The newly interchanged units at the surface of th e crystalline interface are taken into the crystalline region and thus become inaccessibl e to further reaction. These results also show that interchange reactions are confin ed to the amorphous region of the polymer, being severely restricted in the crystalline region. In the literature, a small number of papers explicitly recognize interchange reactions as the dominant mechanism for migr ation of functionality of polyesters and polyamides.31,37 Interchange can result in the formation and swapping of loops and bridges between crystalline regions. Abhiraman and coworkers31 predict that the concentration of loops and bridges, such as those in Figure 1-12, will equal at equilibrium if exchange reactions are taki ng place. They also found th at the jump distance between an amine and an amide in a SSP of Nylon 6,6 was found to be 5.12 , which is in excellent agreement with the 5-6 , which ha s been obtained as the critical reaction distance from classical studies of diffusion-cont rolled reactions of small molecules. This means that the model can be thought of as being analogous to a diffusion-controlled bimolecular system with one or both species diffusing th rough the system, depending on whether or not one or both functional groups in the polymer can undergo exchange reactions. This model seems to be gaining more and more support as time passes. 1.4 Olefin Metathesis Throughout the history of chemistry, any re action that has the ability to form carbon-carbon bonds receives a signif icant amount of attention; olefin metathesis is no
PAGE 41
24 exception. Olefin metathesis is an astoni shing chemical phenomenon, whereby a single chemical reaction has become widely used and st udied in several, very different fields of chemistry in both academic and industrial settin gs. The diversity in application of this reaction has lead to its prevalence in the fields of organometallics,40 organic synthesis,41 and polymer chemistry,42 which have all been thoroughly reviewed.43,44,45 The word, “metathesis,” derived from the Greek words meta (change) and tithemi (place), means an exchange; thus the term “olefin metathes is,” originally coin ed by Calderon in 1967,46 refers to the interchange of carbon atoms (with their substituents ) between a pair of alkene bonds.44 Modern olefin metathesis offers a mild method of forming carboncarbon double bonds. In addition, the double bond offers functionality, which can be subsequently used in a number of other chemistries. There are several distinct modes under which olefin metathesis can operate. These are illustrated in Figure 113. Cross metathesis (CM)47,48,49 is olefin metathesis in its most literal sense, whereby two olefins are exchanging carbon atoms to make new olefins. Here, the selectivity of the reaction is of the utmost importance since a competing reaction is a self-metat hesis of an olefin with itself to give the cross product. Ring-closing metathesis (RCM)50,51,52,53 is just the intramolecu lar version of CM, by which dienes can close to form cyclic struct ures with the release of ethylene. This condensation reaction has found wide use in the synthesis of sophisticated organic molecules. Chiral catalysts have been em ployed for asymmetric RCM, in which one enantiomer of a racemic diene is ring-closed or an achiral diene is ring-closed chiroselectively.53,54 Ring-opening metathesis (ROM)55 is simply the reverse of RCM, whereby an acyclic olefin is introduced to th e cyclic olefin to yield acyclic dienes as
PAGE 42
25 products. Like RCM, this process can also be conducted asymmetrically whereby the added acyclic olefin preferentially adds to one position on the cyclic olefin.55 RCM ROM R + R R1R2+ CM SelfMetathesis R1R2R1R1+ + ROMP n R ADMET R+ n + Figure 1-13. Modes of olefin metathesis in both organic and polymer chemistry There are two modes of polymerization for ol efin metathesis that are governed by very different principles. Ring-ope ning metathesis polymerization (ROMP)42,56,57,58 is a chain polymerization driven thermodynamically by the release of ring strain in a cyclic monomer. This reaction, under the appropri ate conditions with cer tain catalysts, can fulfill the requirement of a “ living” polymerization, producing polymer chains of a welldefined molecular weight.59 Acyclic diene metathesis (ADMET) polymerization60,61,62 is a step-growth equilibrium polycondensation th at is driven by the release of a small molecule (typically ethylene). Essentially, th is is self-metathesis with a diene. During the course of the polymerization, the ethylene condensate is removed, and the equilibrium is shifted toward the polymer. Acyclic diene metathesis can be used to control precise functional group or branch placement along a pol ymer backbone. This has been utilized
PAGE 43
26 in modeling studies of industrial polymer s formed by free-radical polymerization techniques, such as ethy lene/propylene copolymers.63,64,65 1.4.1 Historical Overview of the Mechanism The history of olefin metathesis is quite fascinating.44,45 No one before the 1950s or early 1960s would have ever predicted that a reaction in which a double bond was apparently cleaved and its pieces put back together again woul d be even remotely possible. The first metathesis publication was in 1960, where the ar ticle was on the ringopening polymerization of norbornene, using a catalyst derived from titanium tetrachloride and lithium aluminum tetraalkyls.66 Interestingly, the reaction was termed a “coordination polymerization” in this article. In 1964, the first journal article on a metathesis exchange reaction was published, which reported “olefin disproportionation” of propylene to 2-butene and ethylene.67 Early on, these two fi elds were researched independently from one anot her; the ring-opening polymer izations were typically initiated with Ziegler-Natta type catalysts, such as MoCl5/Et3Al, at or below room temperature, while the exchange reactions we re typically conducted with oxide catalysts, such as MoO3/Al2O3, at elevated temperatures.44 It was not until 1967 that Calderon and coworkers46 recognized that these two reactions we re related, and used the same catalyst (WCl6/EtAlCl2/EtOH in a 1/4/1 ratio) in both reaction types,68,69 beginning the realization of the diverse opportunity of th e olefin metathesis reaction. The determination of the olefin metathesis mechanism was a controversial problem of the late 1960s and the 1970s, whereby several different mechanisms were proposed. The first major mechanism to be proposed was a quasicyclobutane-metal complex mechanism, shown in Figure 1-14. Although th is idea was conceived by Bradshaw and coworkers,70 the principal publication in support of this mechanism was by Calderon and
PAGE 44
27 coworkers in 1968.71 This mechanism was based on the rules of Woodward and Hoffman,72 the bisolefin-metal structure pro posed by Mango and Schachtschneider (shown as starting material in Figure 1-14),73 and the observation of a random distribution of products. The th ree reported steps in the processes were formation of the bisolefin-metal complex, transalkylidenation (Figure 1-14), and a rapid olefin exchange on the metal.71,74 H2C CHR H2C CHR MLnCHR RHC CH2H2C MLnH2C H2CCHR CHR H2C H2CCHR CHR MLnMLnH2CCHR CHR H2C M Sometimes illustrated as: H2CCHR CHR H2C M or Figure 1-14. Early-proposed quasicyclobutan e mechanism for olefin metathesis, which was later found to be incorrect Although later found to be incorrect, this mechanism was proposed because it did explained several traits inherent to the metathesis mechanism. For one, it did explain that the reaction proceeds via scission of the olefin ic bond (transalkidenation), as opposed to a transalkylation, whereby the alkyl substituen ts on the olefinic carbon atoms exchange positions instead. A deuterium study was conducted where 2-butene and 2-butene-d8 were mixed together in the presence of cat alyst, and the only ne w product observed was 2-butene-d4.71 If transalkylation were the primary mechanism, -d2, -d3, -d5, and -d6 species would also have been expecte d. This mechanism also explai ns that a random distribution of products is observed. For example, if 2-pe ntene is mixed with catalyst, the product is 1:2:1 ratio of 2-butene, 2-pentene, and 3-hexe ne, respectively. Similarly, this mechanism also explains the mixture of cis and trans geometric isomers that is observed in products from pure cis / trans starting materials.
PAGE 45
28 Calderon came to these conclusions since he allowed all his metathesis reactions reach equilibrium, which is why he observed a random distribution and the mixture of cis / trans isomers. This mechanism does not account for the appearance of cross products early on in the reaction. In addition, the quasicyclobutane-metal complex is unconventional bonding theory with no prec edent to base this type of reaction intermediate on. In 1972, a second “pair-wise” mechanism was proposed by Grubbs and coworkers with a tetramethylene metallocycle as an intermediate, shown in Figure 1-15.75 This mechanism required the same bisolefin-metal co mplex as a starting point as in the first mechanism. It is interesting that Grubbs, a pioneer in the field of olefin metathesis, proposed this incorrect mechanism. His late r work was a huge cont ribution in providing evidence of the proper mechanism and still to day continues to be a world leader in metathesis. Grubbs had found evidence of a carbon-metal -bond, and with similar -bonded organometallic intermediate s proposed in the work of Katz76 and Eaton,77 this mechanism seemed reasonable. In his work, Grubbs and coworkers claimed that the tetramethylene metallocycle was formed by reacting 1,4-dilithiobutante with WCl6, which would then decompose to form metathesis products. 75 However, no proof that the CHR CHR M CH2CH2M MCH2R R M R R R R RHCCH2M RHCCH2 Figure 1-15. Tetramethylene metallocycle m echanism for olefin metathesis, which was later found to be incorrect
PAGE 46
29 five-membered ring intermediate was actuall y formed was presented, and the rearrangements observed could have occurred after the olefins were formed. Later, Grubbs and coworkers argued that these two “pair-wise” mechanisms (Figures 1-14 and 1-15) w ould actually not give a ra ndom distribution of products.78,79 The case used to demonstrate this was the determination of deuterium content in the ethylene released from mixtures of 1,7-octadiene and 1,7-octadiene-1,1,8,8-d4. Therein, they claimed that a “pair-wise” mechanism would have less ethylene-d2 than in a statistical distribution, which is what is actually observed.78,79 The proper mechanism (Figure 1-16) was first proposed by Herisson and Chauvin80 in 1970 in a publication that was overlooked by most until the mid-1970s.78,79,81 This mechanism is distinctly different from the other two because it involves a metal-carbene complex as the active catalyst, instead of a “pair-wise” interchange. Essential to this mechanism is the presence of a metal carben e (labeled as the initiator or methylene alkylidene in Figure 1-16). Retrospectively, in the days before well-defined catalyses, the carbene must have formed in situ through unknown mechanisms. The metal carbene reacts with olefins to form a metallocyclobutane complex. This complex is similar to the one shown in Figure 1-15 except that it is a four-membered ring instead of five. The metallocyclobutane can then break apart, reforming a metal ca rbene and an olefin. This carbene can then continue in the catalyst cy cle and reacts with other olefins in the solution. This process of the forming and breaking apart of a metallocyclobutane complex is illustrated in each line of the mechanism in Figur e 1-16. In line (a), the initial alkylidene
PAGE 47
30 MCH2initiator CHR' M CHR CH2starting olefin + R'HC MCH2CHR R'HCCHR + MCH2RHCCH2more starting olefin M RHCCH2CH2M CHR CH2CH2new product 1 M CHR CH2CHR M R'HCCHR CH2RHCCHR MCH2new product 2 + + + + (a) (b) (c) Overall Reaction: R'HCCH2R'HCCHR' H2CCH2MCHR R'HCCH2+ + new product 1 new product 2 starting olefin methylene alkylidene more starting olefin Statistical Mixture of: Figure 1-16. Metal-carbene mechanism for olefin metathesis, which is now well established as the correct mechanism (labeled as the initiator) reacts with an ol efin to form a methylene alkylidene. This carbene can then continue along in the catalyst cycle as the catalytically active species. Eventually new products are formed from the pieces of the original olefin, as shown in lines (b) and (c) of Figure 1-16. A random statistical distribution will result at equilibrium since all of these reactions are reversible. Th ere is a small amount of a byproduct (RÂ’HC=CHR) that is formed from the initial alkylidene in line (a). However, since normally only a cataly tic amount of the initiator is added, the amount of this byproduct is usually negligible. In the modern times of well-defined catalysts, this mechanism is represented more appropriately as a catalytic cycle with an initiation period, as is shown for an example of RCM in Figure 1-17. This more clearly illustrates that the methylene carbene is the active specie s that is continually being regenerated as the reaction progresses.
PAGE 48
31 LnMCH2MLnLnM R R initiation Metathesis Catalytic Cycle olefin MLnMLnproduct starting olefin Figure 1-17. Catalytic cycle for olefin metathesis in an example of RCM The metal-carbene mechanism was the only one of the three proposed mechanisms that did not require unusual bonding theory in an intermediate of the mechanism. Also unlike the other two, this mech anism accounted for the appearance of cross products in the early stages of the reaction. Katz and coworkers reexamined the reaction of unsymmetrical acyclic olefins with cyclic olefins and foun d the cross-coupled products appeared before a significant buildup of the scrambled acyclic olef in, which supports the metal-carbene mechanism.81 Grubbs’ deuterium work in refuting a “pair-wise” mechanism also supported this mechanism.78,79 Nearly 30 years later, the metal-carbene mechanism has become well established as th e proper mechanism for olefin metathesis, where well-defined metal-carbene catalysts ha ve found wide use in the diverse field of olefin metathesis. 1.4.2 Development of Well-Defined Catalysts The most significant thing to occur in the fi eld of olefin metathesis in the past 20 years is the development of well-defined catal ysts. These well-defined carbene structures
PAGE 49
32 tend to act more as initiat ors rather than catalysts.44 They offer a better understanding of the mechanism, a much higher degree of contro l, and an allowance to tailor the catalysis to meet the needs of the application. E ach catalyst offers its own advantages and disadvantages, where a varying degree of meta thesis activity, catalyst lifetime, functional group tolerance, and unwanted side reactions ar e possible. A few par ticular catalysts, due to their combination of traits and availa bility, have become tremendously popular, and have found wide use in the field. The first well-defined alkylidene catalysts that were active towards olefin metathesis were developed by Schrock and coworkers; in 1986, tungsten-based complexes were synthesized,82 followed by molybdenum-based complexes the following year.83 These classes of catalysts have recently been reviewed in the literature with a comprehensive list of all the tungsten and molybdenum alkylidene complexes that have been synthesized.84 Common versions of these cat alysts are shown as complex 1 and 2 in Figure 1-18. These high-oxi dation state, 14-electron comp lexes are pseudotetrahedral, electron-deficient, and onl y contain sterically crow ded, covalently bound ligands including the alkylidene. This steric crowding accounts for th e catalystÂ’s stability toward bimolecular decomposition, which was a problem in the development of these catalysts.84 SchrockÂ’s catalysts are extremely active. However, strict anhydrous and inert atmospheric techniques must be employed for their use, and they are not very tolerant of functionality on the olefin substrate. In any catalyst system, functional groups on the substrate or solvent (including oxygen and water) can interfere with the catalytic process.85 If other functionality in the system can bind competitively to the active metal center, the catalyst can be deactivated.
PAGE 50
33 Another possibility is that the functionality can react directly with the metal center, causing decomposition of the catalyst. Mo N Ph O O F3C F3C F3C CF3W N Ph O O F3C F3C F3C CF32 1 Figure 1-18. Common versio n of SchrockÂ’s tungsten ( 1 ) and molybdenum ( 2 ) based olefin metathesis catalysts There were several early reports in the 1960s that the ruthenium complex, RuCl3(hydrate), could perform ROMP in re fluxing ethanol and aqueous emulsion conditions.86,87 The yields for polymer were low, but the fact that metathesis was occurring in a protic environment at all is quite significant. Tabl e 1-1 shows the binding trends that titanium, tungsten, molybdenum, and ruthenium tend to follow. Unlike the other three, ruthenium reacts pr eferentially with olefins over most other species. This gives the ruthenium catalysts st ability towards alcohols, alde hydes, carboxylic acids, and water that the SchrockÂ’s cat alysts do not possess. Regr ettably, ruthenium was not seriously considered as a candidate for welldefined catalysts for more than two decades, most likely because of the discouragingly lo w metathesis activity of ruthenium salts. In the late 1980s, the potential of ruthen ium catalysts were reexamined. It was found that indeed RuCl3(hydrate) did catalyze ROMP; howev er long initiation times were observed in organic solvents.88,89 When strict anhydrous conditions were used, the initiation rates were even slower. Unexpectedly, it was found that ROMP initiated rather quickly in aqueous conditions.90 Water was found, not only to be compatible with the
PAGE 51
34 ruthenium tetrachloride catalyst system, but also it was actually beneficial to the initiation process. Table 1-1. Functional group tolerance of early and late transition metal olefin metathesis catalysts Titanium Acids Alcohols, Water Aldehydes Tungsten Acids Molybdenum Alcohols, Water Acids Aldehydes Alcohols, Water Ruthenium O l e f i n s Acids AldehydesAlcohols, Water KetonesKetones O l e f i n s Aldehydes Esters, Amides O l e f i n s KetonesKetones O l e f i n s Esters, AmidesEsters, AmidesEsters, Amides Increasing Reactivity Since this period, a number of well-de fined ruthenium catalysts have been developed.91 A select group of theses are illust rated in Figure 1-19. The breakthrough came in 1992, when Grubbs and coworkers developed catalyst 3 the first ever metathesis active ruthenium alkylidene.92 They applied a methodology to ruthenium that had been successful in the synthesis of tungsten alkylidenes, whereby 3,3-disubstituted cyclopropenes were used as carbene precursors.93 With this catalyst, norbonene could be polymerized in the presence or absence of water or ethanol. For the first time, a propagating alkylidene coul d actually be observed by 1H NMR during the course of a metathesis polymerization. Although these observations were remarkable, catalyst 3 was limited to ROMP of highly strained cyclic monomers. A trend observed with early transition metal catalysts was that metathesis activity increases w ith electron-withdrawi ng ligands; thus the triphenylphosphine (PPh3) ligands were exchanged with ot her phosphine ligands of lower
PAGE 52
35 basicity in an attempt to improve the catalyst system. After months of testing these new catalysts without success, it was found that one of the largest and most electron-donating phosphines available had a tremendous increase in meta thesis activity.94 This ligand was tricyclohexylphosphine (PCy3), and by replacing the PPh3 ligands to yield catalyst 4 it was learned that the larger and more basic th e phosphine ligands on ruthenium, the higher Ru Ph Cl Cl PCy3PCy3N N Ru Cl Cl PCy3Ph N N Ru Cl Cl PCy3Ph N N Ru Cl Cl O N N Ru Cl Cl O R = Ph 3 R = Cy 4 9 8 7 6 5Ru Cl Cl PR3PR3Ph Ph Figure 1-19. Select well-defined ruthenium metathesis catalyst the metathesis activity, contrary to the thinki ng of the ligands on the SchrockÂ’s catalysts. The -donation of the phosphine ligand to the metal center is thought to promote the formation of the mono-phosphine comple x through the dissociation of the other phosphine ligand and stabilization of a vacant site in the trans -position. This is known to be an important step in the mechanism with ruthenium alkylidene catalyst.95,96,97 When this was initially examined, the dissociati on of a phosphine ligand was surprising since it meant that the mechanism contains electr on-deficient, 14-electron, metallocyclobutane intermediates.95 In addition, the -donation of the remaining phosphine to the metal helps
PAGE 53
36 stabilize these electron-deficient intermed iates. Unlike its predecessors, catalyst 4 could perform ROMP of low ring-st rain monomers, such as cy clopentene, and it became the first ruthenium catalyst that was active toward s acyclic metathesis processes. Due to its robustness in air as a solid and its functional group tolerance, this catalyst could be used in a wide range of syntheti c applications. However, it did require the uses of diphenylcyclopropene in synthesis, which limited its availability. In 1995, a new methodology for synthesizing substituted ruthenium alkylidenes was developed using diazoalkanes as the carbene source.98,99 This methodology led to the synthesis of several more ruthenium alkylidenes, the most active of which was the benzylidene catalyst 5 The activity of this type of cat alyst is highly dependent on the identity of both the Xand Ltype ligands, as well as the alkylidene. Whereas the activity increases with larger, more electron-donating phosphines (PPh3 << P i Pr3 < PCy3), the activity decreases with larger, more el ectron-donating haloge ns (Cl > Br >> I).95 Since incoming olefins may bind trans to a halide, the more el ectron-donating halides will weaken the ruthenium-olefin bond and disfavor olefin coordination, thereby inhibiting the activity of the catalyst. The more electronegative chloride ligands are the most active halogen choice. Also, the alkylidene plays a ro le in the initiation of the catalyst. Alkylsubstituted alkylidenes display more efficient initiation rates than the methylidene version of the catalyst, presumably by assisting in phosphine dissociation. Even faster initiating are ester-substituents; however, these catalysts are not very stable. The benzylidene turns out to be an intermediate case where the phenyl ring is so mewhat electron donating, but it is large enough that it can a ssist in phosphine dissociation. The benzylidene catalyst ( 5 )
PAGE 54
37 can undergo metathesis with ethylene at room temperature to quantitatively form the methylidene complex within minutes.85 Known today as the “first-g eneration Grubbs” catalyst ( 5 ), this 16-electron ruthenium alkylidene is the combination of the bulky, electron-donating PCy3 ligand, the more electron-withdrawing halide ligands, and the fast-initiating benzylidene moiety. It exists in a distorted square pyramidal geomet ry with the alkylidene in the axial position and alternating phosphines and chlori nes in the equato rial positions.99 With its high activity and additional synthese s that have been developed from readily available starting materials,100,101,102 the first-generation Grubbs catalyst has found extensive use in the vast field of olefin metathesis. In 1999, a new generation of ruthenium-base d olefin metathesis catalysts bearing N -heterocyclic carbene (NHC) ligands in pl ace of one of the phosphine ligands was developed, again by Grubbs and coworkers.103 The most notable of these is catalyst 6 which has become known as the “s econd-generation Grubbs” catalyst. N -Heterocyclic carbene ligands are stronger -donors and much less labile than phosphines; therefore these ligands are not able to readily diss ociate from the ruthenium as the phosphine ligands do. Earlier catalysts with two NHC ligands were found to have little improvement over the activity of catalyst 5 .85 However, a mixed-ligand system, as in catalyst 6 was designed in order to utilize the stronger electron-donating ability of the NHC ligand to enhance the dissociation of the more labile phosphine ligand in the trans position. In addition, the ster ic bulk of the NHC ligand, as well as it electron-donating properties more effectively stabilize the electr on-deficient, 14-electron intermediates that promote olefin metathesis. After explori ng different substitution patterns on the NHC
PAGE 55
38 ligand, it was found that mesityl substitution worked well.103,104 It is thought that the mesityl substituents provide the metal with considerable steric protection, which contributes to its hi gh thermal stability. Before this class of cataly sts emerged, it appeared that the gain of functional group tolerance came at the expense of activity. However, catalyst 6 rivals the activity of SchrockÂ’s molybdenum catalyst ( 2 ), but maintains the func tional group tolerance of 5 Catalyst 6 has superior activity to its predecessor ( 5 ), which includes high reaction rates for ROMP of low-strain cyclic olefins, 105 ROMP of the sterically hindered trisubstituted olefins,105 and RCM of sterically demanding dienes to make triand tetrasubstituted olefins.103,104 Also, for the first time, trisubstitute d olefins could be yielded from CM,106 and olefins with a deactivating group di rectly on the double bond could undergo RCM and CM.107 Catalyst 6 has high thermal stability at elev ated temperatures, and has been found to have a longer catalyst lif etime than the parent complex 5 Whereas the catalyst half-life of the methylidene form of the first-generation cataly st is 40 min at 55 C in benzene,108 the second-generation catalyst has a half-lif e of 6 h under the same conditions.103 On the other hand, contrary to what was expected, catalyst 6 does initiates slower, most likely due to sl ower phosphine dissociation.85 These 16-electron ruthenium catalysts coul d conceivably undergo initiation via two general pathways, an associative route or a dissociative route.96,97 Figure 1-20 shows both these general pathways. In an associative pathway, the olefin coordinates to the metal first at a vacant site before the phosphine ligand leaves, thereby making an 18-electron intermediate (A). On th e other hand, in a dissociativ e pathway, the phosphine ligand leaves first, leaving a 14-electon complex (B ) and another vacant site where the olefin
PAGE 56
39 can then coordinate. Regardless of the order in which this occurs, once the phosphine has left, and the olefin has coordinated, the re sulting 16-electron complex (C) can form the metalocyclobutane complex and proceed via the mechanism in Figure 1-16 and 1-17. Recent experimental data96 and density functional calculations97 suggests that it a dissociative pathway where the phosphine departure is the fi rst step of the initiation. Ru Cl Cl PCy3L N N L = PCy3( 6 ) ( 5 ) + olefin olefin + olefin olefin PCy3+ PCy3PCy3+ PCy3R R' Ru PCy3L R' Cl2R Ru L R' Cl2Ru Cl Cl L R' R Ru L R' Cl2(A) (B) (C) (C) Associative Pathway: Dissociative Pathway: or Ru Cl Cl PCy3L R' Figure 1-20. Associative and diss ociative pathways of first( 5 ) and second( 6 ) generation Grubbs catalyst Other notable “second-generation” rutheniu m alkylidene catalysts that have been employed in this body of work are Hoveyda’s catalyst ( 7 ),109 Mol’s catalyst ( 8 ),110,111 and the Hoveyda-Mol hybrid ( 9 ).112 On Hoveyda’s catalyst, th e alkylidene has an isopropyl group in the ortho position of the aromatic ring, on which the oxygen atom coordinates to the ruthenium center in place of the phosphine ligand. Figure 1-21 illustrates how isopropoxystyrene is released and then returned to the ruthenium complex. This in turn allows for the recovery of the catalyst with it s original alkylidene moiety attached to the ruthenium center. The oxygen coordination is quite labile while the alkylidene is attached, which gives it an extremely long cat alyst lifetime in solution. However, in metathesis processes where the isopropox ystyrene dissociates, the catalyst can
PAGE 57
40 decompose if there is nothing to st abilize the 14-electron intermediates.112 This is likely due to molecular isopropoxystyrene not being ve ry labile when not bound to the metal as an alkylidene and therefor e cannot recoordinate with the ease it does when it is attached. Fortunately, most metathesis processes where th is would be used have some functionality on the substrate or on the solv ent (THF has been reported112 to work well) that can O Ru Cl Cl Ru Cl Cl O L N N L = PCy3 or L Ru Cl Cl L O Ru Cl Cl O L ( 7 ) ReleaseReturn Figure 1-21. HoveydaÂ’s catalyst under going RCM, whereby the original oxygencoordinated catalyst is reformed coordinate and fulfill this purpose. MolÂ’s catalyst further increases the steric bulk of the NHC ligand with bulky isopropyl groups on its aromatic rings, which allows for enhanced rates of reaction, particularly at lo wer temperatures. Apparently, the interaction of the isopropyl groups with the alkylidene and/or the chlorides promotes phosphine dissociation. Although this bulkier NHC ligand was orig inally thought to possibly increase catalyst stability,111 it was found to have the opposite effect. At monomer to catalyst loadings of 450: 1, MolÂ’s catalyst has been reported to not go beyond trimerization under ADMET polymerization conditions at temperatures of 45 C and
PAGE 58
41 higher due to cata lyst decomposition.112 Other catalysts have been developed with a varying degree of bulkiness on the NHC ligand. In general, the bulkier the NHC ligand, the more active the catalyst is, but the faster the catalyst decomposes. The Hoveyda-Mol hybrid, developed within the Wagener resear ch group, combines the thinking from the two variations on GrubbsÂ’ second-generation catal yst, gaining the advantages of both in one catalyst. These second-generation catalysts do ha ve a major disadvantage that their processors do not. This is that olefin isomerization has been observed in many metathesis processes.113,114,115,116 Although there are examples of isomerization using earlier metathesis catalysts,113,117 only trace amounts were observed and only after long reaction periods. The extent to that which isomerization occurs with catalyst 6 and other ruthenium catalysts with NHC ligands is much higher. For cases where the metathesis reaction is unfavorable, then isomerization may even be the primary mode of reaction occurring, as was the case in one study,115 where as much as 100 % yield of the isomerization by-product was reported in an attempted RCM reaction to make an eightmembered ring. In an isomerization study fr om within our group that simulated ADMET conditions, Lehman and coworkers found signi ficant isomerization of linear hydrocarbon species occurs at temperatures 50-60 C.113 For one example therein, catalyst 6 was added to the 14-carbon species, 7-tetradecene, at 55 C. If no isomerization were to occur, the expected product w ould be starting material si nce all possible metathesis routes are unproductive. Inst ead, a bell curve of species ranging from nine to twenty carbons was observed by gas chromatogra phy, centered around the 14-carbon species. These other products could only appear if isom erization were occurring. Also reported
PAGE 59
42 therein is that the extent of th is isomerization with catalyst 6 is greatly reduced, but not completely suppressed, at temperatures 30 C and below. Alkene isomerization can occur by two known mechanisms, the metal-hydride addition-elimination mechanism and the -allyl metal-hydride mechanism.118,119 It is not completely clear if the isomerization in metathesis reactions is promoted from the catalyst itself, decomposition products, or impurities in the catalyst synthesis, although even from early on, it has been thought that the most the likely culprit is a ruthenium hydride complex that forms from the decomposition of the catalyst, which would presumably isomerize through the metal-hydride addition -elimination mechanism. Frstner and coworkers reported the isolation of a ruthenium dihydride complex RuCl2(PCy3)2(H)2 during the synthesis of catalyst 8 which could be a cause.110 The key to understanding isomerization may lie in understanding the pathway of catalyst decomposition. Although this pathway is not totally clea r to date, some things are known about the relative rates of decomposition. Firstly, the rate of decompositi on is second order in respect to catalyst concentration, whereby two ruthenium carbene s likely form a bimetallic species as a byproduct. Also, the rate is retarded with the adding of extra phosphine ligand, which implies that the phosphine-dissociated interm ediate is the species that likely couples together. This is consistent with the fact th at catalyst lifetimes are shorter with catalysts that have fast phosphine disso ciates rates. Most recen tly, Grubbs and coworkers may have solved this mystery by isolating structure 10 (Figure 1-22) directly from catalyst 6 under metathesis reaction conditions.120 An in depth isomer ization study with this diruthenium hydride by-product has not yet been conducted; howev er, they did report one
PAGE 60
43 Ru Cl Cl PCy3Ph 10 6 55 oC C6H6+2 CH3PCy3 +Cl-2 N N N N Ru Cl Cl C Ru H N N Mes Mes Cl Figure 1-22. Bimolecular decomposition of catalyst 6 to form diruthenium complex 10 which is likely responsible for isom erization during ol efin metathesis example of isomerization of an olef inic species using this species ( 10 ). At this time, this bimetallic hydride complex appears to be th e most likely culprit responsible for the isomerization that occurs dur ing olefin metathes is of second-generation Grubbs catalyst with similar complexes possible for other s econd-generation catalyst. Also proposed therein is a mechanism of d ecomposition to arrive at this structure (Figure 1-23). 120 This mechanism is not only via a bimolecular pathwa y, but it is also consistent with phosphine dissociation being an important step in the decomposition of the cata lyst. If this new understanding of catalyst decomposition is co rrect, it may now be possible to design a catalyst with the inhibition of this mechanism in mind. Even though olefin metathesis is a latecome r to the modern chemical scene, it has flourished into a widely used reaction in organic and polymer chemistry. Each new catalyst that is developed is another tool available to the metathesis world, each with its own advantages and disadvantages. The more that is understood about the mechanisms that these catalysts undergo, the higher poten tial there will be for synthesizing new catalysts that are tailored to exploit desirable properties.
PAGE 61
44 N N Ru Cl Cl Ru H N N Mes Mes 10 CH3PCy3 +Cl-RuCH2PCy3N N RuCH2Cl Cl PCy3N N Cl2+ RuCH2Cl Cl N N PCy3Ru Cl Cl N N PCy3Ru CH2Cl Cl N N PCy3CH2Cy3P Ru Cl Cl N N + N N Ru Cl Cl C Ru H N N Mes Mes Cl RuCH2N N Cl2Cl Cl CH2N N Ru Cl Ru H N N Mes Mes Cl Cl CH CH2Cy3P 6 Figure 1-23. Mechanism proposed by Grubbs for the decomposition of catalyst 6 1.5 Opportunities of Olefin Met athesis in the Solid State The mission of this work is to demonstr ate that a particular mode of olefin metathesis, ADMET, can operate in the soli d state within the same principles as semicrystalline condensation polymers. Ac yclic diene metathesis, RCM and CM all involve condensation steps, so in principle all of them coul d operate in th e solid state, assuming some degree of amorphous character is present during the c onversion. Of the three, ADMET chemistry leads to macromolecule s, and like the industrial polymers that are typically performed as a solid-state pol ymerization (polyester s, polyamides, and polycarbonates), the equilibrium is shifted towards the polymer by the removal of small
PAGE 62
45 condensate. In most cases of ADMET, the c ondensate that is releas ed is ethylene, as shown in the ADMET mechanism60,121 in Figure 1-24. R n R R R LnRuCH2ethylene is released R RuLnR LnRu RADMET Mechanismpolymer LnRu R' R' initiation olefin R LnRu Figure 1-24. The ADMET mechanism whereby the equilibrium is shifted toward the polymer by the release of ethylene Solid-state ADMET has a distinct advantag e over common solid-state polyester, polyamide, and polycarbonate cond ensation reactions, and that li es in the volatility of the condensate. Table 1-2 compares the condens ate boiling points of typical industrial SSP and ADMET. All other known SSP examples release high boiling condensates such as ethylene glycol (PET), water (Nylon), or phe nol (polycarbonate), all requiring reaction temperatures of greater than 100 C to drive off the condensate. In contrast, the condensate for ADMET boils at –104 C, and therefore can be conducted at significantly lower temperatures. Fortuitously, this al lows for utilization of reaction temperature where metathesis catalyst s are typically run (30-60 C). In addition, lower temperatures are more economically viable and cause fewer side reactions to occur. This process could possibly lead to the gene ration of highly pure, high molecular weight polymers at or near room temperature. Further, the process may offer a new route to the direct preparation of intractable polymers, which pr esently are made via indirect techniques.
PAGE 63
46 Table 1-2. Comparison of condensate boiling points Polymer Type Example Condensate Condensate Boiling Point ADMET PolyoctenyleneEthylene -104 C Polyester PET Ethylene glycol 198 C Polyamide Nylon 6,6 Water 100 C Polycarbonate Lexan Phenol 182 C Since this work is conducted in the solid state, Chapter 2 is a brief explanation of the experimental methods, uncharacteristic of typical organic synthesis, which were developed and used throughout the work. Chap ter 3, which contains most of the work performed for this dissertation, focuses on ADMET polymerization of semicrystalline polymers in the solid state in an effort to mimic the solid-state polycondensation processes conducted in industry. Chapters 4 an d 5 are extensions of this work aimed at more unconventional principles, whereby Ch apter 4 concentrates on solid-state ADMET of rigid-rod polymers and Chapter 5 focuses on attempting ring-closi ng metathesis in the solid state.
PAGE 64
47 CHAPTER 2 EXPERIMENTAL METHODS Since it is the objective of this work to conduct reactions in the solid state, unconventional organic methods ha d to be applied for the solid -state reactions. It is therefore appropriate to devot e a brief chapter on the tec hniques that were used and developed throughout this work. This chapter will descri be the typical solid-state apparatus, the different methods used to make a solid mixture of diene and catalyst, the process of sampling the reaction, and then the characterization techniqu es that were used on the samples throughout this work. 2.1 The Solid-State Apparatus The typical experimental setups were triv ial. The apparatus consisted of equipping a 50 mL three-neck round bottom flask with two vacuum adapters (one requiring a stopcock) and a stopper (all 24/40 joints). The glassware was typically dried overnight in the oven at 80 C. While still hot, the apparatus was assembled whereby the vacuum adapter without a stopcock is connected to a high vacuum/argon line and the vacuum adapter with a stopcock is connected to a bubbler as shown in Figure 2-1. With the stopcock closed, the apparatus is evacuated and then backfilled with argon three times. The line to the bubbler was also purged with argon during the backfilling process. The apparatus was then flame dried and allowed to cool under vacuum. At this point, reactants could be added with positive argon pressure. Reaction temperature was maintained in an oil bath (+/2 C) by one of three temperatur e controllers. The first was a J-KEM Scientific Model 210 with a K-probe (-200 to 1372 C) and a heating coil. The
PAGE 65
48 second was a Barnant Company Model 689-0000 with a T-probe (-200 to 400 C) and a heating coil. The third was a Corning Temperature probe #400188 for Corning PC-420 hot plates. vacuum argon solid bubbler temperature-controlled oil bath Figure 2-1. Drawing of a typi cal apparatus for solid-state reactions (For experiments with argon flow, the stopcock was left open to allow air flow to the bubbler. Likewise, the stopcock was closed fo r reactions subjected to vacuum.) The reactor was then left in tact at the set temperatur e with argon flowing over the solid solution to the bubbler throughout the course of the re action. In some cases vacuum was used with the stopcock closed instead of argon flow. Typical ly these experiments were run for 7 d and sampled periodi cally at specified time intervals. 2.2 Making a Solid Mixture of Reactants Multiple methods of creating a “solid solu tion” of reactants were experimented with. This section aims to convey these methods in detail. The most common method used in this work was an initial polymeri zation in the melt state before subsequent crystallization to a solid mixt ure of prepolymer and catalyst. For polymerizations aimed to start from monomer, as well as reactions aimed to conduct RCM in the solid state, other means of creating a solid mixture were sought out. The freeze-drying of the solid monomer and catalyst together was utilized for these pu rposes using benzene as a
PAGE 66
49 solvent. In later studies of the same t ype, it was found that simply grinding the solid diene and catalyst together in a mortar a nd pestle could produce re action mixtures that achieved similar molecular weight increases. Th is grinding technique was also applied to reactions mixtures of cleaned up (precipita ted to remove catalyst) prepolymers and catalyst for additional polymerization in the solid state. In making up these solid solutions, scal es of 0.2 g to 0.5 g of monomer were typically employed with catalyst loadings typically from 100:1 to 500:1 monomer to catalyst. The details of what was used for each particular experiment including catalyst loading and reaction temperature are specifie d in the results and discussion section as each experiment is conveyed and appears in the results and discussion section of the individual chapters where they are presented. 2.2.1 Melt Polymerization This method most closely simulates indus trial SSP by beginning the polymerization in the melt state with subsequent crystall ization of the growing macromolecule and continued polymerization in the solid state; th erefore this method was most often used in this work. The melt portion of this pro cedure was conducted either at the same temperature as the solid-state portion, or at elevated temperatures to keep the forming polymer in the melt for a period of time. In the former case, the reaction mixture solidified after an amount of time at a constant temperature, after wh ich the solid mixture was first sampled to mark the start of the solid-state reaction, which was then continued at the same temperature. In the latter case, elevated temperatures were needed to keep the monomer and subsequent polymer in the melt for an amount of time before the eventual crystallization by cooling to a more appropriate reaction temperature. In further attempt to simulate the industrial process, the solid was often taken out of the reaction
PAGE 67
50 flask, and broken up into small particles in order to increase surface area to assist in condensate release. This was usually accomp lished by grinding with a mortar and pestle; however, this was sometimes replaced by or complemented with the breaking up of the solid with a spatula. The small particles were then transferred to a fresh apparatus at the desired temperature for continue d reaction in the solid state. 2.2.2 Freeze-Drying For polymerizations aimed to start from monomer and reactions aimed to conduct RCM in the solid state, freeze-d rying of the two reactants together was utilized. Separate solutions of the monomer (or RCM diene) and catalyst were prepared in minimal amounts of benzene and cooled in an ice bat h. Usually one or bot h of these solutions would actually solidify depending on concentr ation. A similar reaction flask to that shown in Figure 2-1 was setup except that both vacuum adapters required a stopcock. One vacuum adapter is connected to a trap, and the whole system is evacuated and flame dried as described above. The two cooled be nzene solutions were then combined in the reaction flask with positive argon pressure, and immediately frozen (freezing point of benzene = -5.5 C) in a dry ice/ace tone bath (-78 C). Once frozen, the stopcock to the trap was closed, and quickly the trap was ev acuated and submerged in a dewer containing liquid nitrogen. At this point the reaction flask was also su bjected to vacuum by opening the stopcock on the vacuum adapter, subliming the benzene from the reaction flask and into the trap with liquid nitrog en as shown in the setup in Figure 2-2. A solid mixture of diene and catalyst results afte r all the benzene has been removed. The amount of benzene collected in the trap was typically very close to the combined amount from the two original solutions (within 1/2 mL for 6 mL of benzene). The complete removal of
PAGE 68
51 benzene and the possible presence of any c onversion from the short time in the liquid state were always checked by NMR spectroscopy. dry ice/acetone bath Vacuum liquid nitrogen Figure 2-2. Freeze-drying apparatu s with liquid nitrogen trap 2.2.3 Grinding Together Diene and Catalyst A solid mixture of diene and catalyst coul d also be made simply by grinding the solid diene and catalyst together with a mortar and pestle. In several cases, the starting materials would stick to the mortar and the pestle; therefore the material was scraped off the sides of the mortar with a spatula after each grinding st ep. This process of grinding and scraping was completed multiple times in order to achieve a more homogeneous mixture. Three grinding and scraping steps were performed for the making of each solid mixture to be consistent throughout these studie s. For the best possibility of obtaining a homogeneous mixture, the two solid starting ma terials were mixed with a spatula prior to
PAGE 69
52 the grinding process to preven t clustering of the catalyst in any one part of the solid mixture. 2.2.4 Regrinding Clean Polymer with Fresh Catalyst This grinding technique was also applied to reaction mixt ures of clean prepolymers and catalyst. Here a prepolymer is first s ynthesized under melt or so lution conditions that will be specified in the individual chapters. The resulting prepolymer was then quenched (see next section) and precipitated (usually in cold methanol) to remove the catalyst. After filtration and drying unde r high vacuum, the prepolymer could then be ground together with fresh catalyst as de scribed in the previous section. 2.3 Sampling of the Solid Mixtures Since solid-state reactions are inherently slow, each experiment was run for an extended period of time (usually 7 d). Th e reaction progress was monitored by opening up the reaction flask with positive argon pressure, and scoping out a small portion of the solid mixture with a spatula. This solid could then be used for molecular weight characterization in order to obser ve the progress of the reaction. In order to prevent further reaction in th e solution states of the characterization solvent, each sample had to be quenched. This was accomplished by the addition of ethyl vinyl ether (EVE), a known inhibitor of ol efin metathesis with the Grubb ruthenium catalysts.94,122,123 These ruthenium-carbene complexes reaction rapidly a nd quantitatively with EVE to form a Fisher carbene as show n in Figure 2-3. The resultant complexes have been shown to be metathesis inactive at standard conditions and only weakly active at elevated temperatures. The addition of EVE was accomp lished in one of two ways. Either it was added via a pre-made quench so lution, or added direc tly to the sample.
PAGE 70
53 O LnRu CH R = H or polymerethyl vinyl ether LnRu O metathesis inactive Fishercarbene + R H2C CH R + Figure 2-3. Reaction of ruth enium catalysts with EVE Another concern was the possibility of radical chemistry occurring while in solution for samples that had to be stored for an extended period of tim e prior to analysis. This was addressed by the addition of a free radical scavenger, 2,6-ditert -butyl-4methylphenol, commonly know as butylated hyd roxy toluene (BHT). The mechanism by which BHT inhibits radical chemistry is s hown in Figure 2-4. A hydrogen radical is transferred from the phenolic oxygen to quench the radical species. The resultant free radical form is stable because the radical moiety is surround ed by the two bulky tert-butyl groups and therefore cannot reac t in typical radical chemistr ies. In experiments where nuclear magnetic resonance (NMR) spectrosc opy was the primary mode of molecular weight determination, BHT was not be used because the OH peak of BHT (5.0 ppm in CDCl3) overlapped with one set of the external olefin peak s from typical ADMET dienes (typical meaning the monomers have long-met hylene spacers between the olefins and any functionality present). In these cases, the need for accurate integration outweighed the need to prevent radical chemistry. Between st oring samples in a refr igerator and taking the measurements as soon as possible, prevention of radical chemistry was sufficiently achieved. OH + R O RH + Figure 2-4. Free radica l scavenging with BHT
PAGE 71
54 For when the 7-d reaction time is completed, quench solutions were prepared with the aim of quenching a 0.5 g batch of polymer which would contain approximately 5 mg of catalyst at 500:1 monomer to catalyst ra tio, with 0.5 mL of solution. After adding EVE and BHT to a volumetric flask, the remaining volume was filled by the GPC solvent. These quench solutions contained EVE to catalyst ratios of 100:1 to 500:1 and BHT loadings of 0.1 % to 5.0 % by weight. When making up individual samples for GPC, the polymer samples (~5 mg) were first quenched by adding 2-5 drops of the pr oper quench solution, and then diluted to the desired concentration with the GPC solvent. Early NMR experiments were quenched in the same manner until it was discovered that the OH-peak was interfering the external olefin integrations. For the NMR experime nts where BHT was excluded, which is the case for the majority of the NMR data presen ted in this work, only EVE was used when quenching the samples. For NMR in chloroform-d (CDCl3), batch quench solutions were prepared simply by adding 1.0 mL of EVE to a 50 g bottle of CDCl3. This is an approximate 100-fold excess of the amount sugg ested of EVE (100:1 EV E to catalyst) to quench a 20 mg sample with the 1.0 mL portion needed for the NMR in a 300:1 monomer to catalyst experiment. Storing thes e solutions in the refrigerator is imperative because EVE decomposes (one by-product obs erved was acetylaldehyde) in the slightly acid chloroform environment at room temper ature. For all other NMR solvents, EVE (2-3 drops) was added directly to the sample with subsequent addition of approximately 1 mL of the NMR solvent. This was f ound to give approximately the same EVE concentrations as the CDCl3 solutions above.
PAGE 72
55 2.4 Characterization Techniques Used on the Samples This section details the general modes of polymer characterization that were used on samples throughout this work. In partic ular, gel permeation chromatography (GPC) and/or NMR end-group analysis were used at the two primary forms of molecular weight determination. Other forms of characteri zation are detailed within the individual chapters. 2.4.1 Gel Permeation Chromatography (GPC) Two different GPC systems were used th roughout this work. The first machine was a modular system with Waters U6K liquid chromatograph equipped with an Eppendorf TC-45 column heater, Polymer Labor atories Ultrastyragel C linear mixed bed columns, a Hewlett Packard HP 1047 refractiv e index detector, and an ABI Analytical Kratos Division Spectroflow 757 ultraviolet detector. On this machine, relative measurements were taken relative in polysty rene in HPLC grade tetrahydrofuran (THF) of cholorform at 1.0 mL/min with sample concentrations of 4.0 to 8.0 mg/mL. The second machine was a Waters GPCV 2K w ith high temperature and light-scattering capabilities. Relative measurements we re taken using HPLC grade THF at 40 C on Waters Styragel HR 5E columns relative to polystyrene standards at 1.0 mL/min. The light-scattering equipment was a Wyatt Techno logies three-angle miniDAWN that was inserted into the heated compartment of the GPCV 2K. On this machine, sample concentrations of approximately 0.5 mg/mL were used. All relative GPC for both machines were calibrated against ten na rrow (PDI > 1.07) polystyrene standards (Polymer Laboratories) ranging in number average molecular weights ranging from 580 g/mol to 275,200 g/mol.
PAGE 73
56 2.4.2 Nuclear Magnetic Resonance (NMR) Spectroscopy The spectra were all taken on a Varian Gemini 300 MHz, a VXR 300 MHz, a Mercury 300 MHz, or Mercur y 300 MHz broadband. In CDCl3, 1H spectra were referenced to TMS as 0.0 ppm and 13C spectra were reference to CDCl3 (t) as 77.23 ppm. In Toluene-d8, 1H spectra were referenced to TMS as 0.0 ppm and 13C spectra were reference to toluene-d8 (s) as 137.86 ppm. In DMSO-d6, 1H spectra were referenced to DMSO-d6 (5) as 2.50 ppm and 13C spectra were reference to DMSO-d6 (7) as 39.51 ppm. In D2O, 1H spectra were referenced to the HOD peak as 4.80 ppm and 13C spectra were not referenced. These reference numbers are all from a Cambridge Isotope Laboratoiries, Inc. NMR solvent data chart. End-group analysis was performed by compar ing the integrations from internal olefin peaks (5.28 to 5.46 ppm) and the exte rnal olefin peaks (5.72 to 5.90 ppm and 4.88 to 5.06 ppm). These chemical shifts are of typical polymer s made via ADMET that have long-methylene spacers between the olefins and any functionality that may be present in the monomer when taken in CDCl3 relative to TMS. For an ADMET polymer with a degree of polymerization (DP) of 1 + n (F igure 2-5), the signal appearing from the internal olefins is from 2n protons per chai n, while the signal appearing from the external olefins is from six protons per chain. Therefor e by comparing the ratio of the integrations from the internal olefins and the sum of the integrations from the external olefins, the values of n and DP can be calculated. A t ypical spectrum of the olefin region along with the calculation of DP is presented in Fi gure 2-6. With the DP, the number average molecular ( Mn) can be calculated simply by multip lying by the DP by the molecular weight of the repeat unit ( Mo). This technique is only useful for molecular weight determination of polymers of relatively low molecular weight (40 kDa is an approximate
PAGE 74
57 upper limit). As the molecular weight is incr eased, the integrations from the end groups eventually disappear and the degree of erro r is too large for accurate molecular weight measurements. R R n R H H H H H H H H 2n H internal olefin 6 H externalolefin DP = 1 + n ADMET int end = 2n 6 Figure 2-5. Relative NMR inte grations of internal and external olefins for NMR endgroup analysis DP = 1 + n = 1 + 3(2.00) int end = 2n 6 = 29.27 4.61 + 10.00 = 2.00 DP = 7.00Mn = Mo x DP R n R H H H H H H H H 4.6129.2710.00 6.00 ppm (t1) 5.505.00 Figure 2-6. Olefin region in the NMR for typical ADMET polymers with long methylene spacers. The calculation to obtain the degree of polymerization shows that this particular sample is only oligomer (DP = 7.00)
PAGE 75
58 CHAPTER 3 SOLID-STATE ADMET OF SE MICRYSTALLINE POLYMERS 3.1 Introduction Although olefin metathesis has been a lat ecomer to the modern chemical scene, it has become a standard chemical conversion in both organic41 and polymer42 chemistry, particularly in the past decade. As such, it is important to understand the different qualities and attributes of each mode of conversion. Of th ese different modes, acyclic diene metathesis (ADMET) has been wi dely employed within our research laboratory.60,61,62 Since the work that led to ADM ET being recognized as an important mode of olefin metathesis,121,124,125,126 a variety of unique pol ymer structures and architectures have been synthesized in our laboratory. These include precisely branched polyethylene,63,64,65,127 silicone-containing polymers,128,129,130,131,132,133,134 metal-containing polymers,135,136 amino acid-containing polymers,137,138,139,140,141 branch copolymers,142,143 and a number of other functionaliti es integrated into the polymer backbone.144,145,146,147,148,149,150,151,152,153,154,155,156,157 Additionally work has been done to better understand the areas where ADMET excels, as well as its limitations.112,113,114,158,159,160,161 Similarly, the work herein intends to expand upon the understanding of this polymerization, aiming to explore the possibi lities of conducting ADMET in the solid state. Before beginning this discussion, it is im portant to understand what is meant by “solid-state polymerization.” Solid-state pol ymerization has tradi tionally been divided into two classes, crystal-to-crystal poly merization and equilibrium condensation polym-
PAGE 76
59 erization.3,4,8 The process that this chapter aims to demonstrate for ADMET polymerization is of the latter category. This type of SSP is a step-growth condensation reaction that is conducted on semicrystalline polymers. This means that the growing polymers have some degree of both amorphous and crystalline charac ter, having both a glass transition temperature ( Tg) and melting transition ( Tm). The polymerization is conducted above the Tg of the polymer, which allows for enhanced molecular mobility in the amorphous region due to a softening with in these non-crystalline areas. It is important to keep in mind that this is termed as a solid-state polymerization in industry even though it is not strictly a crystalline so lid. The amorphous region may be softened above the Tg, however the crystalline region remains intact below the Tm, and the shape of the particle is maintained throughout th e SSP. The enhanced molecular motion within the amorphous region facilitates the release of a small condens ate molecule (ethylene, in the case of ADMET chemistry as shown in Fi gure 1-24), which can be removed at the surface of the solid to drive the increase in molecular weight. Since solid-state condensation is widely used in industry in the syntheses of polyest ers, polyamides, and polycarbonates, many of the experiments he rein were aimed at mimicking these processes to obtain molecular weight increase in the solid state. The one distinct advantage that ADM ET has over these indus trial condensation conversions lies in the volatility of its conde nsate, ethylene. Rem oval of the condensate can be the rate-determining factor in solid -state polycondensation. Table 1-2 compared the condensate boiling points of typical i ndustrial SSPs and ADMET. The industrial examples all require temperatures of greater than 100 C to drive off the condensate, while ethylene can be removed at much lower temperatures (BP = –104 C). This
PAGE 77
60 difference allows for reactions to be performed at temp eratures where metathesis catalysts are typically active (30-60 C). Lower temperatures are also more economically viable since less energy is required for heating. Additional techniques were employed to further aid in condensate removal and thereby mimic the industrial processes. In i ndustry, the growing polym er is subjected to flow of an inert gas or vacuum during the solid-state polymerization. In these experiments, condensate removal was typically ai ded by the flow of argon through the reaction vessel, although vacuum was o ccasionally applied where noted. Also, in some experiments, the prepolymer was ground into smaller particles in a mortar and pestle to mimic the industrial extrusion process, which is done to produce a higher amount of surface area for the condensate to escape. The monomer chosen for the initial wo rk in this area is 1,9-decadiene ( 11 in Figure 3-1) for its simplicity in nature. This “benchmark” hydrocarbon ADMET monomer lacks any polar or protic functionali ty, and thus allows for the as pects of the process to be scrutinized rather than potential complicat ing effects, such as the complexation or reaction with the catalyst. Since this monomer is a liquid, the polymerization to polyocteneylene ( 12 ) is started in the melt state with subsequent crystallization of the resulting polymer, and continue d polymerization in the solid state. These prepolymer-topolymer studies are intended to be a clos e recreation of solid-state condensation processes performed in industry. Monomers with ketone and alcohol func tionality were also shown to undergo polymerization under ty pical ADMET conditions.152,153,154 Alcohol monomer 13 and ketone monomers 15 and 16 which polymerize to 14 17 and 18 respectivel y (Figure
PAGE 78
61 3-1), were synthesized to show that so lid-state ADMET could be conducted with a variety of functionality. These monomers of higher-melting temperatures were initially chosen in order that solid-sta te experiments that start from the monomer and then proceed directly to high polymer could be attempted. There are several examples of solid-state polycondensations that do undergo polymeriz ation from the monomer; however, low conversions have been typically observed.4,5,29,30 Solid mixtures of monomer and catalyst were made by freeze-dying them together, or simply by grinding the solid monomer and catalyst together in a mortar and pestle. O O O x x 6 99 O O O x x 99 n + + + 11 15 x = 9 16 x = 8 19 OH 9 9 OH 9 9 + 12 17 x = 9 18 x = 8 Br Ph Br Ph 20 1314 6 n x x ADMET ADMET ADMET ADMET n n Hydrocarbon Ketone Functionalized Copolymer Alcohol Functionalized Figure 3-1. Monomer used in solid-state ADMET polymerization A polystyrene graft copolymer was also selected to demonstrate solid-state ADMET with different polymer architecture. Macromonomer 19 (Figure 3-1) is essentially a chain of polys tyrene grown from a diene by atom transfer radical polymerization (ATRP),8,162 which could then be polymerized by ADMET.143 This type
PAGE 79
62 of monomer and resulting polymer ( 20 ) would presumably be amorphous in nature since it is essentially in a polystyrene environment. Key factors in any solid-state polycondensat ion are the nature and the lifetime of the catalyst employed, in this case, metathesis catalysts. Three of the most widely used metathesis catalysts are the Grubbs first( 5 )98 and second( 6 )103 generation catalysts and SchrockÂ’s molybdenum catalyst ( 2 ).83 Experiments were conducted comparing the activity of these ca talysts for a variety of conditions. This chapter demonstrates that the two Grubbs catalysts do indeed prom ote solid-state ADMET polymerization. 3.2 Experimental 3.2.1 General All reactions were conducted in flame-dr ied glassware under an argon atmosphere unless otherwise noted. Reagents were purchased from Fische r or Aldrich and were used as received unless otherwise noted. 1,9-Decadiene was distilled from CaH2 before use. Metathesis catalysts used in this chapter, first( 5 )98 and second( 6 )103 generation Grubbs and SchrockÂ’s molybdenum ( 2 ),83 were synthesized by known methods. Instrumentation for GPC and NMR are disc ussed in Chapter 2. Thermal analysis was conducted on a Perkin-Elmer DSC 7 with a Perkin-Elmer CCA 7 temperature regulator. Gas Chromatogra phy was performed on a Shimadzu GC-17A chromatograph equipped with a RTX-5 (Restek Corp.) 15 m column and a HP-5 (H ewlett Packard) 25 m column with FID detection. Elemental anal yses were performed by Atlantic Microlab, Inc (Norcross, GA). Mass spectrometry (MS) was conducted by the University of Florida Mass Spectrometry Services. Melti ng points were performed on a Laboratory Devices Mel-Temp apparatus.
PAGE 80
63 3.2.2 Monomers Syntheses 3.2.2.1 Synthesis of tricos a-1,22-dien-12-ol (13) This monomer was synthesized by one of two Grignard reactions, one using the alkenyl bromide and the other using the alkenyl chloride. Physical attributes of the compound from both routes matched after purification. For the bromide route, a procedure from the literature was modified.154 Magnesium turnings (2.12 g, 87.2 mol) were ground in a mortar and added to a three-neck roundbottom flask equipped with a condenser and an addition funnel. A solution of 11-bromoundec-1-ene ( 22 ) (16 mL, 17 g, 0.074 mol) in dry diethyl ether (60 mL) was added dropwise while warming gently in a water bath (up to 40 C), when a greenish black turbid solution formed. Afte r complete addition (1.5 h), th e solution was refluxed for an additional 2 h, at which time a soluti on of ethyl formate (2.5 mL, 2.3 g, 0.031 mol) in dry ether (10 mL) was added dr opwise (15 min). The resulting solution was refluxed overnight. The reaction mixture was poured onto 200 mL of 3 N HCl/ice, causing a white solid to precipitate out. The layers were separated, and the aqueous portion was extracted with ether (30 mL). The combined organic layers were washed twice with 3 N HCl (50 mL), twic e with water (50 mL), and dried over MgSO4. Upon evaporation of solvent, 10.82 g of th e crude white product was obtained. This material was purified by flash chromatography using 40 % DCM in hexanes, followed by 100 % DCM to yield 6.33 g (61 %) of white tricosa-1,22-dien-12-ol ( 13 ) (MP = 54-57 C). Purification by 10 % et hyl acetate in hexanes has also been reported.154 1H NMR (CDCl3, ppm): = 5.89-5.73 (m, 2H), 5.04-4.89 (m, 4H), 3.69-3.52 (br, 1H), 2.091.98 (q, 4H), 1.62-1.52 (br s, 1H), 1.50-1.18 (br m, 32H). 13C NMR (CDCl3, ppm): =
PAGE 81
64 139.47, 114.31, 72.24, 37.71, 34.03, 29.92, 29.83, 29.76, 29.70, 29.35, 29.15. 25.87. ESI/FT-ICR HRMS: Calcd for C23H44ONa ([M+Na]+): 359.3290; Found: 359.3295. Elemental Anal. calcd for C23H44O: C 82.07, H 13.18, O 4.75; Found: C 82.14, H 13.26, O 4.79. For the chloride route, magnesium turn ings (4.11 g, 87.2 mol) were ground in a mortar and added to a similar setup as a bove. A solution of 11-chloro-undec-1-ene ( 23 ) (23.6 g, 0.125 mol) in dry THF (125 mL) was added dropwise (1 h) and the solution allowed to stir for an additional hour. The temperature of the reaction flask was reduced to – 78 C with a dry ice/acetone bath and et hyl formate (4.6 mL, 4.2 g, 0.057 mol) was added dropwise, after which the resultant so lution was allowed to reflux overnight. The reaction was quenched with 3 N HCl and the a queous portion was extracted with ether. The combined organic layers were washed tw ice with 3 N HCl, tw ice with water, and dried over MgSO4. After evaporation of the solvent, 10.8 g (96 %) of the crude white product was obtained the resultan t crude product was recrystal lized in acet one twice to obtain white tricosa-1,22-dien-12-ol ( 13 ), having the same physical properties as that described above. 3.2.2.2 Synthesis of tricos a-1,22-dien-12-one (15) This synthesis was performed by Paul va n Gervan within our research group. A slurry of pyridinium chlorochromate (2.69 g, 12.5 mmol) in 10 mL of dry dichloromethane (DCM) was prepared in a 100 mL three-neck round-bottom flask equipped with two stoppers and an addition f unnel, to which a solution of diene alcohol 17 (3.26 g, 9.69 mmol) in 20 mL of dry DCM wa s added dropwise. After the addition (2 h), the black slurry was stirred overnight at room temperature. Diethyl ether
PAGE 82
65 (25 mL) was added and the resulting turbid so lution was decanted. This addition of ether and decanting was repeated twice. The et her/DCM solution was passed over a plug of silica and rotovaped to yield a white solid with a green tint. The solid was dissolved in 35 mL of hexanes, washed three times with 25 mL of water, dried over MgSO4, and subsequently evaporated under reduced pressure to yield 2.62 g (80.1 %) of white tricosa1,22-dien-12-one ( 15 ) (MP = 58-59 C). 1H NMR (CDCl3, ppm): = 5.90-5.73 (m, 2H), 5.04-4.88 (m, 4H), 2.38 (t, 4H, J = 7.2 Hz), 2.03 (q, 4H, J = 7.2 Hz), 1.55 (p, 4H, J = 7.2 Hz), 1.44-1.17 (br, 24H). 13C NMR (CDCl3, ppm): = 211.96, 139.44, 114.33, 43.04, 34.02, 29.62, 29.48, 29.32, 29.13, 24.01. HRMS: Calcd for C23H42O: 334.3236; Found: 334.3240. Elemental Anal. calcd for C23H42O: C 82.57, H 12.65; Found: C 80.88, H 12.59. 3.2.2.3 Synthesis of heneicos a-1,20-dien-11-one (16) Synthesis was performed using a pro cedure published in the literature,139 which was modified from the original sy nthesis by Zantour and coworkers.163 10-Undecenoyl chloride (21.5 mL, 0.100 mol, 20.3 g) and dry diethyl ether (150 mL) were added to a 500 mL three-neck round-bottom flask equipped with a reflux condenser and an addition funnel, and then cooled to 0 C. Triethyl amine (25.1 mL, 0.180 mol, 18.2 g) was added dropwise, instantly precipitating white trie thylammonium chloride salt. The reaction mixture was allowed to warm to room temper ature and stir for 24 h. After salt filtration and ether evaporation, the intermediate lactone was mixed with 100 mL of a 2.0 M solution of NaOH (8.00 g, 0.20 mol) a nd refluxed overnight under argon. The reaction mixture was acidified with 200 mL of 1.0 M HCl and extract ed three times with ether. The combined ether layers were washed twice with 1.0 M HCl, twice with
PAGE 83
66 saturated NaHCO3, and once with brine, dried over MgSO4, and recrystallized in MeOH to yield 9.60 g (62.7 %) of pur e heneicosa-1,20-dien-11-one ( 16 ) (MP = 52-55 C). 1H NMR (CDCl3, ppm): = 5.90-5.72 (m, 2H), 5.04-4.87 (m, 4H), 2.38 (t, 4H, J = 7.2 Hz), 2.02 (q, 4H, J = 7.2 Hz), 1.56 (m, 4H), 1.35 (m, 4H), 1.32-1.20 (br, 16H). 13C NMR (CDCl3, ppm): = 211.93, 139.42, 114.35, 43.03, 34.01, 29.58, 29.52, 26.47, 29.29, 29.11, 24.09. HRMS: Calcd for C21H38O: 306.2923; Found: 306.2932. Elemental Anal. calcd for C21H38O: C 82.28, H 12.50, O 5.22; Found: C 81.96, H 12.74, O 4.99. 3.2.2.4 Synthesis of the polysty rene-graft macromonomer (19) This monomer was prepared via the literature method.143 The atom transfer radical polymerization (ATRP) step is reported here while precursor steps to achieve initiator 24 are reported therein. Styrene was washed with NaHCO3 and deionized water to remove inhibitors and distilled over CaH2 under reduced pressure befo re use. Diphenyl ether was recrystallized in 95 % ethanol in water, melted, dried over CaCl2, and then fractionally distilled before use. Copper(I) bromide (0.43 g, 3.0 mmol) and 2,2 Â’-bipyridine (0.94 g, 6.0 mmol) were weighed out and placed into a flame-dried, argon-purged, 50-mL Schlenk tube with a stir bar. Diphenyl et her (5 mL) was added via a syringe, and the mixture was subsequently degassed with three freeze-pump-thaw cycles. Initiator 24 (0.75 g, 1.5 mmol) and styrene (9.37 g, 90 mm ol) were added via a syringe to the aforementioned solution. The resultant mi xture was stirred for 3 h at 110 C while maintaining a slight argon flow. The so lution changed from brown to green upon heating. The reaction mixture was cooled to room temperature and quenched with THF (60 mL), filtered through a neutral bed of alum ina to remove residual catalyst, and filtrate was concentrated to 30 mL. Macromonomer 19 was obtained by prec ipitation into cold
PAGE 84
67 methanol. Finally, this solution was filtered, and the white solid product 19 was dried under reduced pressure. Integrations in NMR showed that approximately 60-65 polystyrene units were on each chain. 1H NMR (CDCl3, ppm): = 7.23-6.870 (br, 203H), 6.87-6.28 (br, 130H), 5.905.70 (m, 2H), 5.08-4.90 (m, 4H), 4.50 (t, 1H), 3.51 (d, dH), 2.04 (q, 12H), 1.98-1.68 (br, 56H), 1.68-1.08 (br m, 176H), 1.02-0.80 (br m, 8H). 13C NMR (CDCl3, ppm): = 177.81, 157.16, 1 45.24, 139.14, 129.66, 127.91, 125.60, 123.13, 118.82, 114.12, 66.81, 41.82, 40.33, 37.02, 33.80, 31.26, 29.96, 29.59, 29.14, 28.93, 26.67. NMR end-group analysis: Mn = 7400. GPC: Mn = 6800, Mw = 9200, PDI = 1.35. 3.2.3 Polymer Syntheses Methods of creating a solid solution, runni ng a solid-state reaction, modes of quenching, and types of molecula r weight characterization ar e discussed in Chapter 2. The details of what was used for each partic ular experiment including catalyst loading and reaction temperature are specified in th e results and discussi on section as each experiment is conveyed. In a typical experiment, each sample was analyzed for molecular weight by GPC and/or NMR end-group analysis. Any additional characterization will appear in the re sults and discussion as well. Full characterization has not been included since the polymers synthesized herein have been synthesized before, and only trends in molecular weight increa se in the solid state were desired. 3.2.4 Other Molecules Synthesized These compounds ( 21 23 ) are all known and the physi cal properties of the product match those of Aldrich and have not been included. These procedures appear in the literature from Baughman and coworkers164 within our group.
PAGE 85
68 3.2.4.1 Synthesis of undec-10-en-1-ol (21) Zinc undecylenate (250 g, 0.579 mol) was added over 30 min via a powder funnel to a stirred slurry of lithium aluminum hydride (LAH) (25.0 g, 0.659 mol) in dry THF (400 mL) in a 5 L round bottom flask at 0 C. After the addition, the solution was allowed to warm to room temperature over 1 h while stirring. Th e reaction was quenched via addition of deionized wate r (25 mL), 15 % (w/v) NaOH (25 mL), and more water (75 mL) waiting approximately 15 min between add itions. The solution was allowed to stir until cool, and the reaction mix appeared as a white slurry. All precipitate was filtered, and the solution was concentrated to a turbid oil. The crude mixture was dissolved in ether and stirred over MgSO4 (15 g) for 45 min. The solution was filtered and concentrated to afford 85.8 g (87 %) of undec-10-en-1-ol ( 21 ) as a colorless oil (97 % pure by GC). 3.2.4.2 Synthesis of 11-bromo-undec-1-ene (22) A solution of undec-10-en-1-ol ( 21 ) (150 g, 0.881 mol) and carbon tetratbromide (323 g, 0.976 mol) in dichloromethane (800 mL) was prepared in a 2 L flask and cooled to 0 C. Triphenyl phosphine (255 g, 0.976 mol) was added via a powder funnel in portions over 30 min with vigorous stir ring. Upon addition of the phosphine, the colorless solution turned a pale brown color a nd was stirred for an a dditional 2 h at room temperature. The mixture was concentrated to a brown oil and quick ly added to stirring hexanes (1 L). The white precipitate wa s filtered, and the remaining solution was concentrated and fractionally distilled yielding 195 g ( 95 %) of 11-bromo-undec-1-ene ( 22 ) as a colorless liquid. In addition to Al drich, the physical prope rties matched those published previously in our group.154
PAGE 86
69 3.2.4.3 Synthesis of 11-chloro-undec-1-ene (23) In an argon-purged 3 L round-bottom flask, distilled thionyl ch loride (259 g, 2.17 mol) was added dropwise over 1 h via cannula into a solution of undec-10-en-1-ol ( 21 ) (200 g, 1.28 mol) in pyridine (50 mL). Upon complete addition, the reaction was heated to 50 C for 2 h, cooled, and quenched via the addition of water (300 mL) and diethyl ether (300 mL) letting st ir for 1 h. The remaining mixture was extracted, and the organic phase was washed twice with saturated NaHCO3 (150 mL) and once with distilled water (100 mL). The solution was dried over MgSO4, concentrated, and vacuum distilled to yield 169 g (70 %) of 11-chloro-undec-1-ene ( 23 ) as a colorless oil. 3.3 Results and Discussion Much of the data that appears in this section also appears in the literature;165 however, several experiments that do not appear in the origin al communication have been included, along with a more thorou gh discussion of the results. 3.3.1 Prepolymer-to-Polymer Experiments Solid-state experiments that start with partially polymerized material closely mimic industrial solid-state polycondensati on. Monomers of varying functionality and architecture (Figure 3-1) were shown to undergo solid-state ADMET in a manner similar to the SSPs conducted in industry. 3.3.1.1 Solid-state polymeriza tion of 1,9-decadiene This initial work with this simple hydrocarbon monomer demonstrated that ADMET polymerization could be conducted in the solid state. It was found that by starting the polymerization in the liquid stat e at a constant temperature, the growing polymer would crystallize, and th en continue molecular weight in crease in the solid state.
PAGE 87
70 Data for catalysts 5 and 6 at 500:1 monomer to catalyst ra tios is displayed in Table 3-1 for 30 and 45 C. Table 3-1. Results of solid-sta te polymerizations of monomer 11 with periodic sampling Cat Mon : Cat Ratio Temp ( C) Number Average Molecular Weightsa at Specified Reaction Times 3 h 6 h 12 h 24 h 48 h 72 h 96 h 168 h 5 500 : 1 30 2,8003,700 6,600 9,400 9,900 19,000 20,00021,000 3 h 6 h 12 h 30 h 48 h 72 h 107 h 5 500 : 1 45 5,4006,400 7,200 16,00030,000 30,000 31,000 3 h 6 h 12 h 24 h 48 h 119 h 6 500 : 1 30 2,2005,800 7,000 7,800 7,900 8,800 3 h 6 h 12 h 24 h 48 h 72 h 96 h 168 h 6 500 : 1 45 8,90012,00012,00015,00018,00018,000 21,00021,000 a Mn determined by GPC in chloroform re lative to polystyrene standards. These results demonstrate unequivocally that solid-state polycondensation is occurring at temperatures as low as 30 C, with the possibility of the reaction occurring at even lower temperatures. Th is is the first ex ample of a solid-state polycondensation being conducted as low as room temperatur e, as all other exam ples are conducted at elevated temperatures. Figure 3-2 shows th e GPC traces for polymerizations done with first-generation Grubbs catalyst at 45 C. As the polymerization proceeds, the number average molecular weight increases from 5,400 to 31,000 g/mol, all while in the solid state. It should be noted th at all the molecular weights repor ted herein are well within the range normally expected for equilibrium step -growth polymerization chemistry. Table 3-2 illustrates that the molecular weights obser ved in this work are within the range of useful molecular weights for typical step growth polymers, which exhibit useful properties with number average molecular weights of less than 25,000.166
PAGE 88
71 Figure 3-2. Gel Permeation Chromatography traces of the SSP of 1,9-decadiene with first-generation Grubbs at 45 C sampled periodically (The molecular weights listed at 3, 30, and 107 h are number av erage molecular we ights determined by GPC in chloroform relative to polystyrene standards.) Table 3-2. Useful molecular weight s of typical step-growth polymers Polyesters: Mn ~ 15,000 to 25,000 Polyamides: Mn ~ 15,000 to 30,000 Polycarbonates: Mn ~ 12,000 to 30,000 ADMET Polymers: Mn ~ 10,000 to 30,000 PolymerUseful Molecular Weights Both catalysts 5 and 6 promote solid-state ADMET polymerization. Firstgeneration Grubbs ( 5 ) was found to be the more active of the two, largely due to its higher reactivity at lower temperatures, reac hing number average molecular weights of 30,000 g/mol in 3 d at 45 C. Second-generation Grubbs ( 6 ) is slower, but seems to continue to promote polymerizat ion at later stages of the reaction. This observation may be the result of the longer lifetime of complex 6 as compared with the first-generation
PAGE 89
72 catalyst. This is similar to what is seen in the literature in solution. The half-life of second-generation Grubbs is found to be 6 h at 55 C in benzene,108 compared to 40 min for first-generation Grubbs;103 however, we suspect that the lifetimes of these complexes are longer in the solid state compared to those in solution. Experiments with SchrockÂ’s molybdenum catalyst ( 2 ) were conducted at 45 C with a 1500:1 monomer to cataly st ratio. Since this catalyst is much more sensitive to oxygen, the reaction mixture had to be starte d in a glove box under an argon atmosphere and then transferred to an outside line to achieve argon flow. Two experiments were run side by side for 75 h, one with sampling in the glove box, and one without. For both experiments, the molecular weight of the first sampling period equaled the molecular weight of the final sampling at 75 h (Table 3-3). Any molecula r weight fluctuation observed is within the error of the technique. The likely re ason that SchrockÂ’s catalyst was not observed to be active is deco mposition of the catalyst. Although two experiments cannot rule out th e possibility that catalyst 2 can be active in the solid state, further work herein was focused on the activity of th e Grubbs catalysts 5 and 6 in the solid state. Table 3-3. Results of solid-state polymerizations using SchrockÂ’s catalyst 2 and monomer 11 Method of Sampling Mon : Cat Ratio Temp ( C) Number Average Molecular Weightsa at Specified Reaction Times Outside 3 h 25 h 50 h 75 h Glove box 1500 : 1 45 1,600 1,500 1,500 1,600 Inside 8 h 27 h 75 h Glove box 1500 : 1 45 1,800 1,800 1,800 a Mn determined by GPC in chloroform re lative to polystyrene standards. Since this is a solid-state reaction where mo lecular mobility is relatively low, it was thought that increasing the catal yst loading would potentially provide a route to higher
PAGE 90
73 conversion. The cata lyst loading was increased five-f old (100:1 monomer to catalyst ratio) with the second-generation catalyst ( 6 ) and is compared with the corresponding 500:1 reactions at 30 and 45 C in Table 3-4. At both temperatures, comparable molecular weights were achieved with both catalyst loadings, although, the observed molecular weights of the experiments with hi gher catalyst loadings actually tended to be slightly lower. These experiments suggest that either catalyst load ing, or anywhere in between, is acceptable to achieve molecular wei ght increase in the solid state, but higher catalyst loadings do not promote the polymerization to higher conversion. Table 3-4. Catalyst loadi ng comparison using monomer 11 and catalyst 6 Cat Mon : Cat Ratio Temp ( C) Number Average Molecular Weightsa at Specified Reaction Times 3 h 6 h 12 h 24 h 48 h 119 h 6 500 : 1 30 2,2005,800 7,0007,8007,900 8,800 5 h 8 h 24 h 48 h 72 h 120 h 6 100 : 1 30 2,8005,800 6,7006,800 6,700 8,100 3 h 6 h 12 h 24 h 48 h 72 h 96 h 168 h 6 500 : 1 45 8,90012,00012,00015,00018,00018,000 21,00021,000 3 h 6 h 24 h 48 h 72 h 96 h 168 h 6 100 : 1 45 4,0005,900 10,00014,000 15,000 16,00018,000 a Mn determined by GPC in chloroform re lative to polystyrene standards. When comparing polymerizations of the sa me catalyst at different temperatures, higher temperatures achieved higher molecular weights (Table 3-1). For first-generation Grubbs ( 5 ) at 30 C, a number average molecula r weight of 19,000 g/mol was achieved in 3 d, while at 45 C, it reached 30,000 g/mol in the same amount of time. A similar effect was found with secondgeneration Grubbs catalyst ( 6 ). At 30 C, the number average molecular weight does not exceed 10,000 g/mol in 5 d, while reaching approximately 21,000 g/mol at 45 C in the same amount of time.
PAGE 91
74 Similar experiments to these were conducted at 60 C where ultra-high molecular weights were achieved with catalyst 6 However, all but one of these experiments did not appear to be a solid-state reaction as the reaction mixtures were viscous melts. The presence of liquid properties was tested simply by making a hole in the viscous mixture with a spatula; hours later the hole would be filled because the material had exhibited flow properties. In one expe riment (line 1 in Table 3-5), however, the reaction mixture did indeed solidified before the first sampling time at 3 h. Several attempts to reproduce this experiment and decipher why this solid ification occurred were unsuccessful. All experiments conducted at 60 C are presented in Table 3-5. The most striking piece of data from these experiments is that the molecular weight achieved by the one solid-state reaction (line 1) was 122,000 g/mol after four days. After 7 d, however, the molecular weight decreased to 79,000 g/mol, which is what is observed in typical solid-state polyc ondensation with extended reac tion periods (Figure 1-11). Since a prior attempts at 60 C started outside the glove box with catalysts 6 and 5 (line 2 and 5, respectively) had been observed to be a viscous melt, this solidification was initially thought to be from less catalys t decomposition through starting the reaction inside the glove box. To test this theory, another experiment star ting in the glove box (line 3) was conducted in the exact same manner to yield a viscous melt during polymerization. This suggested that it was no t lack of catalyst d ecomposition that caused it to solidify. However, ultra high molecular weight was observed in this experiment as well (131,000 g/mol after 7 d). This high molecu lar weight is more easily reasoned for a melt reaction. These polymerizations are neat reactions (no solvent), therefore when in the melt state, the olefinic species, bot h monomer and growing polymer, are at the
PAGE 92
75Table 3-5. Experiments conducted at 60 C using monomer 11 Line Cat Mon : Cat Ratio Started in Glove Box Number Average Molecular Weightsa at S p ecified Reaction Times State Observe d 3 h 6 h 12 h 24 h 48 h 72 h 96 h 168 h 1 6 500 : 1 yes 12,00011,000 17,00016,00014,00037,000122,00079,000 Solidb 3 h 6 h 12 h 24 h 48 h 72 h 96 h 168 h 2 6 500 : 1 no 17,00022,000 44,00047,00066,00080,00080,00098,000 Melt 3 h 6 h 12 h 24 h 48 h 72 h 96 h 168 h 3 6 500 : 1 yes 14,00024,000 23,00026,00037,00043,00052,000131,000 Melt 3 h 6 h 12 h 24 h 48 h 72 h 96 h 168 h 4 6 100 : 1 no 15,00020,000 20,00029,00028,00035,00034,00035,000 Melt 3 h 6 h 12 h 24 h 48 h 72 h 96 h 168 h 5 5 500 : 1 no 12,00015,000 14,00015,00014,00014,00013,00015,000 Melt a Mn determined by GPC in chloroform (line 1-3, 5) or THF (line 4) relative to polystyrene standards. b This state was not reproducibly observed.
PAGE 93
76 highest concentration possible. This, combin ed with the molecular mobility of a liquid state, gives the olefin end-groups the best chance to find each other and achieve this ultrahigh molecular weight. Another possibility that was tested was that the solidification came from an error in weighing out the catal yst causing a higher cata lyst loading. An experiment (line 4) with a higher catalyst loading (100: 1) was run, again yielding a viscous melt. Although this is not a solid-state reaction, it is interest ing to note that this higher catalyst loading did ind eed yield significantly lower molecular weights. Even though the cause of the solidification (line 1) is not clear, the fact remains that it did occur, and the molecular weight did incr ease from 12,000 g/mol to 122,000 g/mol in the solid state. This is a 10-fold increase to yield ultra high molecular weight polymer. The condensate in solid-state polycondensations is rem oved in industry by a flow of inert gas or by vacuum, whereby the choi ce is usually based on convenience. For the majority of these experiments, flow of argon through the reaction vessel was used. However, the use of vacuum to remove the condensate was tested. Table 3-6 compares the number average molecular weig hts of these two methods at 45 C with catalyst 6 For the vacuum experiment, a round-bottom fl ask was equipped with a bath condenser, and a Schlenk adapter inside a glove box. Th e monomer and catalyst were added to the flask in the glove box, and then the apparatus was tr ansferred to an outside line with the Schlenk stopcock closed. After filling the ba th condenser with dry ice/isopropanol and submerging the reaction flask at 45 C, the apparatus was subjected to intermittent vacuum until the reaction mixture solidified. Sampling for this experiment was conducted inside the glove box. The molecular we ight achieved at the end of the vacuum experiment was less than half that of the ar gon experiment. Because of this observation
PAGE 94
77 and complications with volatility of monomer 11 argon flow is the method of choice for these experiments. Table 3-6. Comparison of argon and vacuum experiments using monomer 11 and catalyst 6 Condensate Removal Mon : Cat Ratio Temp ( C) Number Average Molecular Weightsa at Specified Reaction Times 3 h 6 h 12 h 30 h 48 h 72 h 107 h argon 500 : 1 45 5,4006,400 7,200 16,00030,000 30,000 31,000 5 h 27 h 72 h 168 h vacuum 500 : 1 45 2,300 12,000 13,000 13,000 a Mn determined by GPC in chloroform (argon) or THF (vacuum) relative to polystyrene standards. Experiments also were conducted to increa se the surface area of the solid mixture in order to facilitate escape of the condensat e from the surface of the solid. This also further simulates the industrial process whereby the prepolymer is extruded into small pellets for the same reason. One such experime nt is illustrated in Figure 3-3. 1,9-Decadiene ( 11 ) was polymerized in the melt with second-generation Grubbs at 70 C for 10 min and then quickly crystallized by cooli ng to room temperature. The resulting low molecular weight semicrystalline polymer ( Mn = 1,300 g/mol by NMR) was 6 6 x 6 y 10 min at 70 oC (melt) 7 Days at 45 oC (solid state) Mn = 1,300 g/mol Mn = 23,000 g/mol 66 11 Figure 3-3. Prepolymerization in the melt of 1,9-decadiene, followed by polymerization in the solid state using second-gene ration Grubbs ruthenium catalyst ( 6 ) (Number average molecular weights ( Mn) were obtained by NMR end-group analysis in CDCl3.) ground into a fine powder with a chilled mortar and pestle. Th is finely divided solid was brought to 45 C and sampled over time as before, showing that Mn steadily increases
PAGE 95
78 over the course of a week, reaching 23,000 g/mol by NMR. This represents an approximate 20-fold increase in molecular weight, all while in the solid state. 3.3.1.2 Alcohol and ketone functionalized monomers Alcohol monomer 13 and ketone monomers 15 and 16 were synthesized in the routes designated in Figures 3-4 and 3-5. Alcohol 13 was synthesized via one of two Grignard reactions fr om alkenyl bromide 22 or alkenyl chloride 23 (Figure 3-4).154 These alkenyl halides metathesis synthons, as well as alkenyl alcohol synthon 21 were synthesized via routes laid out within our group from r eadily available, inexpensive starting materials.164 Ketone 15 can be easily made by oxidation of alcohol 13 with pyridinium chlorochromate (P CC) (Figure 3-4). Ketone 16 was synthesized via a completely different route from the acyl ch loride (Figure 3-5). This synthesis was OH 9 9 O 9 9 OH 9 Br 9 Cl 9 8 O O 2 Zn2+O HOEt 2122 2313 15 O HOEt Mg0Mg0CBr4, PPh3DCM SOCl2pyridine LAH THF PCC DCM Figure 3-4. Synthetic route to alcohol 13 and ketone 15 modified within our group139 from a procedure first performed by Zantour and coworkers.163 This procedure is a one-pot synthe sis from 10-undecenoyl chloride, which happens to be only acyl chloride -olefin that is currently readily available and inexpensive. Since ketones 15 and 16 only differ in one methylene group on each side of the diene, they were considered equivalent in terms of these solid-state experiments.
PAGE 96
79 Monomer 16 became the monomer of choice for the majority of the ketone experiments because of simplicity of the one-pot synthesis. In fact, ketone 15 was only used in the monomer-to-polymer experiments (Section 3.3.2). O 8 8 8 O Cl 7 C O 7 C O C O NEt3NaOH, H2O 7 22 77 OH OOH CO277 O OOH 77 OH 16 taut. taut. Figure 3-5. Mechanism of synthesis of ketone 16 (This mechanism has been published in references [139] and [163].) Experiments with monomers 13 and 16 similar to those described above for 1,9-decadiene demonstrated that a variety of functionalities are amenable to solid-state ADMET polycondensation. For the Grubbs catalysts 5 and 6 functionality of this type is not an issue since ruthenium reacts preferentia lly with olefins before alcohol and ketone functionalities (Table 1-1).85 Since higher molecular weight increase can be achieved with increasing surface area, as was observe d with 1,9-decadiene in Figure 3-3, all of these alcohol and ketone experiments that start from a prepolymer were ground in a mortar and pestle in air to yield a fine powder before begi nning the solid-state portion of the experiment. The melting points of polymers 14 and 18 was significantly higher than that of polymer 12 so the grinding could be done at r oom temperature (not with a chiller
PAGE 97
80 mortar). Another difference from the 1,9-de cadiene experiments is that these experiments were conducted partially in the melt state at elevated temperatur es to make a solid solution since these monomers were solids at room temperature. The prepolymer synthesized in the melt portion of the reaction has been labeled as the “start” sample, as all subsequent molecular weight increase was in the solid state. Examples of these experiments are presented in Table 3-7 for alcohol 13 and Table 3-8 for ketone 16 Table 3-7. Examples of solid-state polymerizations of alcohol monomer 13 Melt Reaction Cat Mon : Cat Ratio Time Temp SSP Temp Number Average Molecular Weightsaat Specified Reaction Times Start 1 d 4 d 7 d 6 300 : 1 5 min 80 C 60 C 2,100 5,200 5,600 6,500 Start 4 d 7 d 6 300 : 1 5 min 115 C60 C 8,200 9,200 9,900 a Mn determined by NMR end-group analysis in toluene-d8 at 80 C. The two examples of polymerizations from alcohol 13 that appear in Table 3-7 have different melt reacti on temperatures. At 80 C, a number average molecular weight of 2,100 g/mol by NMR end-group analysis was achieved in 5 min. During seven days of reacting in the solid state, the molecular we ight increases to 6,500 g/mol, a three-fold increase while in the solid state. At 115 C, a higher molecular weight of 8,200 g/mol was achieved in the sample amount of time in the melt phase. Alt hough steady increase in molecular weight is observed, the change in molecular weight in the solid state to 9,900 g/mol after 7 d was less significant. This may be due to catalyst decomposition at the higher reaction temperatures du ring the melt portion of the reaction. For the five ketone experiments in Ta ble 3-8, the higher temperature (110-115 C) were actually required to maintain the pol ymer as viscous liquid during the melt portion of the reaction. All of these examples showed a trend of molecular weight increase in the
PAGE 98
81 solid state. Using the se cond-generation catalyst ( 6 ), a molecular weight of 8,700 g/mol was observed by GPC after 3 min at 110 C. The molecular weight then increased to 18,000 g/mol over the next seven days, more than doubling while in the solid state. At 10 min in melt reaction, a similar trend was observed, increasing from 6,500 g/mol to 11,000 g/mol while in the solid state. Unlike the alcohol ( 13 ), the longer time in the melt did not yield the higher molecular weight. However it did lead to a slightly less significant conversion as was observed in Table 3-7. Table 3-8. Examples of solid-state polymerizations of ketone monomer 16 Melt Reaction Cat Mon : Cat Ratio Time Temp SSP Temp Number Average Molecular Weightsa at Specified Reaction Times Start 3 h 3 d 5 d 7 d 6 500 : 1 3 min 110 C80 C 8,700 12,000 14,000 18,00018,000 Start 6 h 1 d 3 d 5 d 7 d 6 500 : 1 10 min 115 C80 C 6,500 9,000 9,700 9,500 10,00011,000 Start 3 h 1 d 3 d 5 d 7 d 5 300 : 1 5 min 110 C80 C 1,100 1,400 3,000 4,600 4,700 5,300 Start 1 d 3 d 5 d 7 d 5 300 : 1 10 min 110 C80 C 1,100 3,200 4,400 5,200 4,900 Start 3 d 5 d 7 d 5 300 : 1 45 min 115 C80 C 1,300 2,600 3,400 3,500 a Mn determined by GPC in chloroform relative to polystyrene standards for experiments with catalyst 6 and by NMR end-group analysis in CDCl3 for experiments with catalyst 5 Polymerizations with catalyst 5 did not achieve as high mo lecular weights as with catalyst 6 although the same trends of molecular weight incr ease were also observed. All three experiments with catalyst 5 shown in Figure 3-8 achieved only low number average molecular weights of 1,100-1,300 g/ mol by NMR end-group an alysis during the melt portion of the experiment. The most si gnificant molecular weight increase here was again the reaction that spent the shortest time in the melt phase. At 5 min in the melt, the ending molecular weight after 7 d was 5,300 g/mol. This number got progressively
PAGE 99
82 smaller with more time in the melt stat e reaching only 4,900 g/mol for 10 min and 3,500 g/mol after 45 min. It would appear that the longer the mixture is kept in the melt, the more that the catalyst decomposes. Theref ore, there is less catalyst left in the solid solution for the polymerization in solid state, and a lower molecular weight is achieved. The last entry in Table 3-8 with an exte nded reaction period in the melt (45 min) was actually conducted in hopes of achieving a high er molecular weight of the prepolymer. Instead, even after this extended reaction period, the molecular weight remained quite low (1,300 g/mol), and it consequently had this negative impact on molecular weight increase in the solid state due to extensive catalyst decomposition. When comparing reactions with catalysts 5 and 6 in Table 3-8, there are two differences to note. The mode of molecular weight determination is different, as well as a slight difference in catalyst loading. Although these differen ces are present, they do not invalidate a comparison of the results. Th e second-generation Grubbs catalyst performed better in these experiments si nce it could better withstand the higher temper atures during the melt portion of the reaction, and therefore more catalyst remained active for the solidstate portion of the experiment. Experiments were conducted with purified pr epolymer as well (Table 3-9). After a short reaction in the melt state, cooling, a nd grinding into a fine powder, the portion of the prepolymer was quenched with EVE, dissolv ed in THF, and then precipitated in cold methanol. After being dried under high vacu um overnight, the resulting white polymer was reground together with fresh catalyst (300 :1 repeat unit to catalyst ratio) and run aside the typical, unprecipitated material cont aining the original catalyst left over from the melt portion of the reaction. The amount of fresh catalyst used in the regrinding was
PAGE 100
83 not based on the moles of polymer, but rather the moles of repeat uni t within the polymer, calculated from the weight of the solid prep olymer and the number average molecular weight obtained from NMR e nd-group analysis. This allows for a more accurate comparison between the precipitated and unpr ecipitated control e xperiments since the original catalyst ratios for the melt reactions were based on moles of monomer. As in any polymer precipitation, the purified produ ct was somewhat fractionated by the washing away of lower molecular weight portions to yield slig htly higher number average molecular weights. In both sets of e xperiments in Table 3-9, one for 5 min in the melt and one for 10 min, this apparent increas e in the number average molecular weight during precipitation process was from 1,100 g/mol to 1,500 g/mol. Table 3-9. Precipitation effects on the solid-state polymerizations with catalyst 5 and ketone 16 Melt Reaction Preca Mon : Cat Ratio Time Temp SSP Temp Number Average Molecular Weightsb at Specified Reaction Times Start3 h 1 d 3 d 5 d 7 d no 300 : 1 5 min 110 C80 C 1,1001,400 3,000 4,600 4,700 5,300 Start 7 d yes 300 : 1 5 min 110 C80 C 1,500 9,600 Start 1 d 3 d 5 d 7 d no 300 : 1 10 min 110 C80 C 1,100 3,200 4,400 5,200 4,900 Start 1 d 3 d 5 d 7 d yes 300 : 1 10 min 110 C80 C 1,500 4,700 6,300 7,000 7,500 a Precipitation of the pre polymer after the melt portion of the polymerization. b Mn determined by NMR end-group analysis in CDCl3. Not only were trends in molecular weight increase observed for the precipitated experiments, but also this increase wa s more pronounced than in the unprecipitated control experiments. At 5 min in the me lt, a precipitated portion, reground with fresh catalyst and allowed to polymerize for 7 d at 80 C in the solid state, achieved molecular weight of 9,600 g/mol. This is compar ed to the 5,300 g/mol achieved in the
PAGE 101
84 unprecipitated experiment. At 10 min, 7,500 g/ mol was achieved in the precipitated experiment and 4,900 g/mol with the unprecipitated. The obvious conclusion here is that the fresh catalyst that has not been subjected to the elevated temperatures and no catalyst decomposition has occurred, which allows for a faster increase in molecular weight. Although, it is interesting that the final mo lecular weights from the two precipitation experiments are different considering that they both have the same starting molecular weight (1,500 g/mol). Potent ially, not all of the cata lyst was removed during the precipitation. If this were the case, a highe r molecular weight would be expected from the experiment that spent less time in the melt, which is what was observed. However, this is unlikely since the material is treated with excess EVE to ensure that the ruthenium species are quenched and not metathesis active ev en if they are present in the precipitated prepolymer sample. 3.3.1.3 Polystryene-graft macromonomers Other polymer architectures operate in the solid state in a similar manner. Macromonomer 19 ( Mn = 6,700 g/mol by GPC relative to polystyrene standards in chloroform) was synthesized143 via atom transfer radica l polymerization (ATRP) of initiator 24 to create an amorphous “graft” of pol ystyrene at the center of a long-chain diene monomer (Figure 3-6). This macromonomer undergoes ADMET solid-state polymerization at 90 C when combined with catalyst 6 (20:1 macromonomer to catalyst) achieving a number average molecu lar weight of 13,000 g/mol over 24 h. This reaction occurs in spite of the large pol ystyrene branch attached to the diene. With only two olefins per macromonomer, it is significant that sufficient molecular mobility exists in the solid state to allow pol ycondensation metathesis chemistry to occur.
PAGE 102
85 In fact, even small changes in temperature have a large effect on this molecular mobility. At 75 C, an almost negligible amount of solid-state polymeriz ation occurs with macromonomer 19 over a 2-week period, whereas at 90 C the number average molecular weight doubles in 24 h. ATRP styrene O O 99 O O 99 + 19 Br Ph Br Ph 20 x x ADMET n O O Br 99 24 Mn = 6,700 g/mol Mn = 13,000 g/mol Figure 3-6. Atom transfer radical polymerization (ATRP) of initiator 24 to macromonomer 19 with subsequent ADMET to polymer 20 (Macromonomer 19 was synthesized analogous to the procedure in reference [143].) 3.3.2 Monomer-to-Polymer Experiments Solid-state polycondensations have two dis tinct starting points. While by far the most common starting point is a prepolymer fo rmed in a liquid state, there are examples where the starting point is the monomer.4 In these cases, the reactivity of the monomer is determined to a considerable degree by the ch aracteristics of the parent crystal. The monomer crystal must have an acceptable arrangement of molecules, in which the reacting functional groups are close enough in proximity and have the proper orientation to allow the reaction to o ccur in the solid state. In order to test the possibility of ADM ET achieving polymer directly from the monomer in the solid state, a monomer that was solid at room temperature had to be synthesized. Ketone 15 (MP = 58-59 C) was synthesized by van Gervan within our group to successfully fulfill this objective. A sample of this monomer (0.250 g scale) and catalyst 6 (100:1 monomer to catalyst ratio) were freeze-dried together using benzene
PAGE 103
86 (3 mL total) as a solvent. The resultant solid solution was tested to reveal that no reaction (degree of polymerization = 1.0) had occurred in the solutions during the freeze-drying process by the absence of internal olefin in the NMR spectra. The reaction flask was then placed in an oil bath at 45 C and allowed to react for 1 w eek. The results from this experiment are shown in Tabl e 3-10. The degree of polymer ization (DP) is presented instead of the number average molecular weig ht because high molecular weight was not achieved. Table 3-10. Results from the monomer-t o-polymer freeze-drying experiment using ketone 15 Cat Mon : Cat Ratio Temp ( C) Degree of polymerizationa at Specified Reaction Times Start 20 h 48 h 120 h 144 h 168 h 6 100 : 1 45 1.0 2.3 2.4 2.4 2.5 2.6 a DP determined by NMR end-group analysis in CDCl3. Although these are not high molecular weights (DP of 2.6 corresponds to a Mn of 730 g/mol), it is significant that any convers ion was observed since it occurred in the solid state. In fact, a degree of polymeri zation of 2.6 corresponds to a 62 % conversion of functionality. This is not a terrible yield for typical organic reactions, and it is remarkable that this has occurred in spite of the restraints imposed in the solid state. The nature of step-growth polymers may be the inhibitor of the synthesis of large polymer chains here since significant molecular weig ht is only achieved after high conversion (> 95 %). Once small conversion (DP > 2.0) was achieved, the observed melting point of the solid solution in creased to ~110 C, and the reaction temperat ure could subsequently be increased while the material maintains the appearance of a solid, although when taking a melting point of the solid, around 52 C, which coincides with the melting point of
PAGE 104
87 ketone 16 (MP = 52-55 C), there is a very slig ht change that is bare ly visible to the eye, but it is not a change that causes liquid to pool at the bottom of the capillary. The material appears to be a solid before and afte r this event that is most likely the melting of unreacted crystalline portions of the monomer. These experiments were conducted with ketone 16 and are shown in Table 3-11, where two methods of preparing solid solutions of monomer and catalyst are compared. The two methods are freeze-drying using benzene as a solvent and grinding together monomer and catalyst with a mortar and pestle. After 24 h, the melting point of the mixture was approximately 110 C, which is just about the temperature needed to keep the polymer in the melt during the prepolymerto-polymer experiments. At this point, the temperature was increased from 45 C to 80 C to see if higher convers ions could be achieve. Table 3-11. Results from the monome r-to-polymer experiments using ketone 16 and catalyst 6 (100:1 monomer to catalyst ratio) Method of Solid Solution 1st Temp ( C) 2nd Temp ( C) Temp Switch Degree of polymerizationa at Specified Reaction Times Start 3 h 8 h 24 h 27 h 3 d 5 d 7 d Freezedrying 45 80 24 h 1.0 2.0 2.1 2.2 5.9 8.9 9.7 10.4 Start 3 h 7 h 24 h 27 h 4 d 7 d Grinding 45 80 24 h 1.0 1.2 1.8 2.2 7.0 9.9 11.3 a DP determined by NMR end-group analysis in CDCl3. In both experiments, the degrees of polymerization achieved after 24 h at 45 C (both had DP = 2.2) were comparable to that achieved with the fr eeze-drying of ketone 15 (DP = 2.3 at 20 h) in Table 3-10 for approxima tely the same amount of time. This is in accord with ketones 15 and 16 acting in a similar manner. Once the temperature was increased to 80 C, which is where the prepolym er-to-polymer experiments were conducted, the degree of polymerization increas ed significantly for both methods. After
PAGE 105
88 only 3 h at the higher temperature (27 h), the freeze-drying experiment was at a DP of 5.9 and the grinding experiments was at 7.0. After 7 d, the DP reached 10.4 with freezedrying and 11.3 with grinding. These corres pond to 90-91 % conversion in the solid state. Although these conversions are very close with bot h methods, it would appear that the grinding method performed slightly better. When taking a melting point of th e solid solution after 24 h at 45 C, the sample does not melt until ~110 C. However, this change around 52 C, which is most likely the melting of unreacted crysta lline monomer, may be causi ng the monomer to act as a temporary solvent or plasticizer at this te mperature for the polymerization to occur, followed by resolidification upon its reaction. This could be why a significant jump in conversion is observed after 3 h at 80 C. This theory of partial melting and resolidification will be discussed further in Chapter 5 in the discussion of solid-state RCM. In this freeze-drying experiment in Ta ble 3-11, the solid solution was broken up after the freeze-drying process to increase surface area. This apparently is an important factor in achieving this c onversion. Whereas the grindi ng experiments already have a high surface area, the solid, freeze-dried solutio n needs to be broken up with a spatula for comparable conversions. In another freeze-dry ing experiment run prior to this one with the same temperature increase after 24 h, the sample was partially broken up by routine sampling whereby the rest of the material wa s left as a large chunk. Later samples from that experiment taken from th e two different portions indeed had significant differences in conversion. At the end of 7 d, a DP of 8.3 was achieved from the broken up portion, where only 3.8 was observed from the larg e, unbroken chunk. For this reason, all
PAGE 106
89 subsequent freeze-drying experiments were broken up, such as the one presented in Table 3-11. Potentially, the increase in Table 3-10 co uld have been higher in the latter stages of the experiment if the am ount of surface area had been in creased by this procedure. 3.3.3 Functional-Group Diffusion While physical diffusion of chain segments and ethylene from within the solid to the surface is undoubtedly occurring, the situatio n is more complex than just that, for the increase in molecular weight cannot be rationalized by physical diffusion alone. Semicrystalline condensation polymers are t hought to undergo solid-state polymerization due to enhanced mobility that exists within the amorphous region when above the glass transition temperature. However, the motions of amorphous chain segments in semicrystalline polymers are restrained by being anchored in the crystalline region.31 A more recent model suggests that migration of functionality can be explained by interchange reactions, a phenomenon termed "functional-group diffusion" that has been decribed by Abhiraman and coworkers31 and Lenz and Schuler.38 This is accomplished through a series of exchange reactions of act ive end groups reacting with functionality in the backbone of the polymer, as was illustrated for polyesters in Figure 1-12. We believe functional-group diffusion is e ffectively aiding in the migration of functionality through the amorphous phase of the polymer during solid-state ADMET chemistry. In addition, the diffusion of et hylene to the surface of the solid can be complemented by the same mechanism. Figur e 3-7 illustrates how solid-state functionalgroup diffusion operates in the case of metath esis polycondensation chemistry. Pathway A shows how the ruthenium carbene migrates within the polymer sample simply as a result of metathesis inte rchange reactions, which are congruent with the proposed mechanism for ADMET.60,121 Reactions of this type do not lead to an increase in
PAGE 107
90 molecular weight, but rather to chain length re distribution that result s in a most probable LnRu LnRu RuLnLnRu RuLnRuLnRuLnLnRu LnRu LnRu RuLnRuLnLnRu LnRuLnRuPathway A Pathway B Pathway C Figure 3-7. The ADMET mechanism at wo rk in the amorphous region of a semicrystalline solid (Pathway A shows the diffusion of a ruthenium carbene by interchange reactions without a net change in molecular weight. Pathway B illustrates an increase in molecular weight by formation of ethylene. Pathway C illustrates how diffusion of et hylene to the surface is aided by the same mechanism.) molecular weight distribution ( Mw/ Mn = 2.0). Nevertheless, functional groups are “diffusing” through the solid, which assist s end-group encounters. Pathway B shows how ethylene is generated thereby increasing mo lecular weight. Unless the ethylene is removed from the reaction mixture, the mol ecular weight increase is thwarted by the occurrence of the reverse reac tion. While diffusion of ethylene is the primary mechanism of ethylene migration to the surface, an illust ration of how ethylene transport is aided by
PAGE 108
91 the metathesis mechanism is shown in Pathway C. As ethylene is removed at the surface of the particle, the equilibrium is shifted towards the polymer, and higher molecular weight is achieved. 3.3.4 Thermal Analysis When thermal analysis with differential scanning calorimetry (DSC) is normally performed, multiple scans are conducted to erase thermal history. However, for these experiments, the thermal history is importa nt because it dictates the environment in which the solid-state polymerization occurs. Although multiple scans were gathered, it is for this reason that all the DSC thermographs included here are the first scan before thermal history is erased. The objective of these experiments was to learn more about the environment in which the polymerizations take place. A DS C scan illustrating the typical environment that the 1,9-decadiene ( 11 ) experiments were subjected to is shown in Figure 3-8. This particular sample is the first scan of the st arting sample (1,300 g/mol) of the experiment which saw an approximate 20-fold increase in molecular weight (Figure 3-3). Since the reaction temperature of that experiment was 45 C, the solid-state portion of the experiment was started within the tail of the melting endotherm. Although these polymers visibly appear as solid materials a nd maintain their shape throughout the course of the 7-d solid-state reacti on, there may be partial melting of the crystalline region that enhance molecular mobility in the sample that encourages the molecular weight increase. This phenomenon was also noted for ketone monomer 16 Figure 3-9 shows a typical scan of this polymer. This particular sample was precipitated after 10 min in the melt at 110 C (start sample shown on th e forth line of Tabl e 3-9). The melting
PAGE 109
92 endotherm seems to have three different overl apping melt transitions, one of which is just below the reaction temperature of 80 C. The molecular wei ght increased from 1,500 g/mol to 7,500 g/mol during the course of a 7d experiment. Again this increase may be enhanced from starting the polymerization within the tail of the melting endotherm. -80 -4004080120160200 50 40 30 20 Temp (oC)Heat Flow Endo Up (mW)6 n Mn = 1,300 Figure 3-8. Typical DSC showing the enviro nment at the start of a 1,9-decadiene ( 11 ) experiment (This particular sample is the start sample of the experiment shown in Figure 3-3. The dotted line indicates the reaction temperature (45 C) of that experiment.) -80 -4004080120160200 50 40 30 20 Temp (oC)Heat Flow Endo Up (mW)O 8 8 n Mn = 1,500 Figure 3-9. A DSC of a precipit ated prepolymer from monomer 16 (The dotted line indicates the reacti on temperature (80 C) of that experiment.) Although this tailing effect may indeed be a factor, when the material is traced at the end of the experiment, the melting endothe rm has moved to a higher temperature to a place where the reaction temper ature lies completely out of the endotherm tail. Figure 3-10 illustrates this for the 1,9decadiene discussed in Figure s 3-3 and 3-8. After 7 d, the
PAGE 110
93 melting point endotherm peak not only resided at a higher te mperature (10 C difference at the highest point of the peaks), it was more concise without much of a large -4004080120160200 80 60 40 20 Temp (oC)Heat Flow Endo Up (mW)100 6 n Mn = 1,300 Start Mn = 23,000 7 Days Figure 3-10. Both the DSC scans of th e starting and 7-d samples from the 1,9-decadiene experiment in Figure 3-3 (The dotted line indicates the reaction temperature (45 C) of that experiment.) tailing effect. As the polymerization pro ceeds, the sample could be becoming more uniformly crystalline. One explanation fo r this is a phenomenon, termed SSP-induced crystallization, that has been observed in ot her step-growth solid-sta te polycondensations, whereby crystallizable sequences formed th rough exchange reactions are taken into the crystalline region.4 These chains are consequently denied the enhanced mobility of the amorphous region and become inaccessible to fu rther reaction. The more uniform crystal causes the melt endotherm to be higher, and in this case, the increase observed was to a point beyond any tail on the endo therm. If SSP-induced crystallization is occurring here, this could also explai n the slower molecular weight in creases at longer reaction times, where a plateau of molecular weight was reache d for the majority of these experiments. The high level of crystallinity would inhib it release of the condensate, and molecular weight increase would be halted.
PAGE 111
94 In order to test th e role of the endotherm tail, a series of experiments were conducted with ketone 16 A batch of low molecula r weight ketone polymer 18 was prepared in a chloroform solution with the first-generation catalyst 5 The resulting polymer was twice dissolved in toluene and precipitated in methanol, the second time washing several times with boiling methanol to remove any tail. The DSC of this oligomer is shown in Figure 3-11. Even though this has a similar number average molecular weight (1,200 g/mol) to the sample shown in Figure 3-9 (1,500 g/mol), the endotherm is higher and there is no tail present within the reaction temperatures. Solidstate polymerization temperatures of 50 and 80 C were chosen for different reasons. Since there is typically so me degree hysteresis in DSC scans, a temperature 40 C lower than the melting endotherm, 50 C, was chosen to ensure that the reaction temperature is -80 -4004080120160200 50 40 30 20 Temp (oC)Heat Flow Endo Up (mW)O 8 8 n Mn = 1,200 60 Figure 3-11. A DCS trace of ketone polymer 18 that was synthesized in solution, precipitated, and washed with boiling methanol to eliminate the tail on the melt endotherm (The dotted lines indicate the two reaction temperatures (50 and 80 C) that experi ments were conducted.) completely out of the tail. The other temperature, 80 C, was chosen for comparative reasons since it was this temperature that the majority of the SSPs with ketone 16 were conducted. The prepolymer was ground together with either catalyst 5 or 6 (300:1 monomer to catalyst ratio) and placed in the typical solid-state reac tion apparatus at one
PAGE 112
95 of the two reaction temperatures. The result s of these four experiments are shown in Table 3-12. Table 3-12. Results of SSPs st arting from a prepolymer of 18 without a tail on the melting endotherm that extends into the reaction temperature Cat. Mon : Cat Ratio Temp ( C) Number Average Molecular Weightsa at Specified Reaction Times Start 1 d 4 d 7 d 5 300 : 1 50 1,200 1,200 1,300 1,300 Start 1 d 4 d 7 d 5 300 : 1 80 1,200 1,300 1,400 1,600 Start 1 d 4 d 7 d 6 300 : 1 50 1,200 1,200 1,200 1,200 Start 1 d 4 d 7 d 6 300 : 1 80 1,200 1,300 1,400 1,400 a Mn determined by NMR end-group analysis in Toluene-d8 at 80 C. Although there were slight changes in mo lecular weight, no significant increases from the 1,200 g/mol prepolymer was observed in any of these experiments by NMR end-group analysis. The highest molecula r weight achieved wa s only 1,600 g/mol with catalyst 5 at 80 C after 7 d in the solid state. With catalyst 6 at 80 C, only 1,400 g/mol was achieved. When compared to the preci pitated ketone experiments with catalyst 5 in Table 3-9, which achieved 9,600 g/mol (5 min in melt) and 7,500 g/mol (10 min in melt) after 7 d at the same temperature, these mo lecular weights achieved here are almost negligible. At 50 C, the increase in molecular weight was even smaller. The ending molecular weights were 1,300 g/mol and 1,200 g/mol with 5 and 6 respectively. Even though there was virtually no change in molecu lar weight, these sli ght increases (with 6 by adding another significant figure, the increase was from 1,180 to 1,220 g/mol) do indicate that metathesis is still able to occur at a temperatur e well below the melt endotherm. However, it does indicate that the ta iling effect does play a crucial role in the molecular weight increase in these experiments.
PAGE 113
96 This tailing effect may even be a factor in the polyester, polyamide, and polycarbonate industries, for they are also st ep-growth polymers that have a distribution of molecular weights within the prepolymer (at equilibrium, Mw/ Mn = 2.0) and do not undergo a purification step between the melt and solid phase portions of the process. The reactions in the solid st ate are typically run 5-40 C below the semicrystalline melting point,4 which where these experiments were conducted as well. After all, the experiments presented herein that did cont ain a tail on the endotherm were meant to mimic the processes conducted in industry. Thes e prepolymers did in fact appear as solid material throughout the course of the reaction without partial melting, which is what is observed in industry since part ial melting would cause sticking of the polymer chips and siege the reactor. In industry, if there is any tail on the melting e ndotherm, the solid-state polymerizations may very well be started within it. The presence of a distribution of molecu lar weights may be the key to achieving significant molecular weight increase. Th e smaller molecules may be acting as plasticizers for the larger chains, thereby e nhancing the molecular mobility within the material. This was congruent with what wa s seen in many of the GPC data that was obtained, which all had polydisp ersities close to 2.0. Many of the GPC traces for the 1,9-decadiene experiments, particularly with catalyst 6 in fact had a small shoulder on the low molecular weight side of the peak that was included in the integration of the data presented above. This implies that low molecular weight oligomer is present in the sample, which is what may be acting as a pl asticizer. In additi on, there was an even lower molecular weight peak ( Mp = ~1300 g/mol) that was excluded. Figure 3-12 is the GPC curve of the 24-h sample using catalyst 6 at 30 C (third line of Table 3-1),
PAGE 114
97 illustrating the manner in which the peak was integrated to obtain the molecular weight data ( Mn = 7,800 g/mol). Although this exclude d low molecular weight peak was originally thought to be attr ibuted to the presence of low molecular weight organic molecules such as EVE, and BHT, or disso lve water and oxygen in the GPC solvent, it may have in fact also have been low mol ecular weight oligomer or unreacted monomer, having the effect of a plasticizer. 51015 Retention Time (min)Refractive Index SignalMp = ~1300 Mn = 7,800 Peak with shoulder 6 n Figure 3-12. Typical GPC curv e of polymer from monomer 11 and catalyst 6 (This particular sample is after 24 h at 45 C.) 3.4 Conclusions This work unequivocally proves that solid-state ADMET pol ycondensation with semicrystalline polymers is a reality. Monomers of varying functionali ty and architecture were shown to polymerize under conditions meant to mimic those of industrial SSP whereby the polymerization is started in the me lt and then continued in the solid state. Both first-generation and s econd-generation Grubbs cataly sts were shown to achieve molecular weight increase in the solid stat e. In cases where the reaction mixture was exposed to elevated temperatur es for a period of time, th e second-generation catalyst performed better since it is more robust at these temperatures. Different temperatures, methods of condensate removal, catalyst loadings, and methods of solid solution
PAGE 115
98 preparation were explored. Breaking up of th e prepolymer to increase surface area was found to yield more significant molecular weight increases. Experiments starting from the monomer did not yield high molecular weight polymer, although 62 % conversi on of functionality was obs erved at a temperature 5-10 C below the melting point of the monomer. This is not a terrible yield for typical organic reactions, and it is remarkable that th is has occurred in sp ite of the restraints imposed in the solid state. Even highe r conversions (91 %) were observed with a temperature increase to a point higher than the melting point of the monomer. However this observation may be due to the melting of unreacted monomer and causing it to act as a temporary solvent to the polymerization. We believe that “functionalgroup diffusion” is playing a role in the molecular weight increases since the observed increase cannot be rationalized by physical diffusion alone. Reactive functionality can react with olefin functionality in the polymer backbone in a series of exchange reactions that result in the migration of functionality. These types of interchange reactions fac ilitate the meeting of two en d groups for the production of ethylene, causing a molecular weight increase. The diffusion of ethylene to the surface of the particle where it can be removed can be aided by the same series of exchange reactions. Thermal analysis showed that these polymer izations were typically being started in the tail of the melting endotherm. Upon reactio n in the solid state, however, the melting endotherm shifted to a higher temperature and became more concise. We believe this is likely due to SSP-induced crysta llization, whereby long-chai n segments are taken into crystalline sections and beco me unreactive toward further polymerization. Experiments
PAGE 116
99 conducted on prepolymers with this tail re moved demonstrated an almost negligible amount of conversion, which implies that the endotherm tail may be important to achieving molecular weight in the solid state. We believe this effect to be due to low molecular weight oligomer and unreacted monomer acting as plasticizers while the polymerization occurs. This may also be in factor in industrial SSPs since the industrial reactions were conducted in a similar manner to these experiments at temperatures 5-40 C below the melting point with a purifica tion step in between the melt and solidstate portions of the polymerization.
PAGE 117
100 CHAPTER 4 SOLID-STATE ADMET OF RIGID-ROD POLYMERS 4.1 Introduction An obvious objective of solid-state polymer ization is to synt hesize intractable macromolecules that are unattainable by any other method, intractable meaning insoluble in all typical organic solvents. These pol ymers are either inaccessible by any synthetic means or are only available via indirect routes, usually through a soluble polymer precursors. Most intractable polymers are of a rigid-rod nature, whereby the macromolecules are not afforded the same fl exibility as the semicrystalline polymers discussed in Chapter 3. The bonding in these polymers is locked in place, usually by sp2 and sp bonded centers. This inhibits an y bond rotation along the polymer backbone. Long-chain segments of polymer extend linear ly until there is a kink in the polymer backbone, where the direction th e polymer chain follows is ch anged. This could occur, for example, as a defect in the polymer structure, such as a cis double bond in place of a trans These linear long-chain segments often st ack together to form crystalline domains, which is the underlying cause of their insolub ility. Many rigid-rod polymers are aromatic in nature, where -stacking of the aromatic rings enhance the intermolecular chain stacking of these linear segments. Assuming that some chain mobility is required for a solid-state polycondensation to occur, chai n-end encounters will be significantly hampered by the rigidity of these polymer chains and it may be difficult to conduct this type of polymerization in the solid state.
PAGE 118
101 Conjugated polymers are often rigid in na ture. The area of conjugated polymers has drawn a vast amount of interest in the past few decades.167,168,169,170,171,172,173,174 In 1977, Shirakawa and coworkers demonstrat ed that a film of polyacetylene175,176 could achieve high level of conductivity when doped.177 Since then, a multitude of research groups have engaged in work within this ar ea, highlighting the importance of the field with the awarding of the Nobel prize in chemistry to Shirakawa,178 MacDiarmid,179 and Heeger180 in 2000. Applications include batteri es, fuel cells, light-emitting diodes (LEDÂ’s), electronics, electrochromic displays and devices, electr oluminescent devices, and chemical sensors, to name just a few. Conducting polymers have the capability of being used in applications where conductive inorganic mate rials have typically been used. Some advantages over inorganic se miconductors and metals are that conducting polymers are lightweight, flexible, less toxic, and generally less expensive. However, the most important advantage is the extent to which the properties of the material (optical, electrical, mechanical, etc.) can be tailored. Organic materials have a synthetic flexibility that allows for the engineering of materials to meet specific application needs. For this work, which aimed to conduct solid-state ADMET on rigid-rod polymers, poly( p -phenylene vinylene) (PPV) polymers were chosen because monomers in the form of a divinylbenzenes (DVB) possess the di ene functionality needed for an ADMET polymerization. Unsubstituted PPV is only available from indirect synthesis through soluble polymer precursors, which was first done by Wessling and coworkers.181,182,183 A typical route by which this is performed is shown in Figure 4-1. This procedure, however, yields polymer with a high defect content, which arises from incomplete conversion of the elimination st ep of the soluble pr ecursor where the vinlyene groups are
PAGE 119
102 formed. While this material is useful for many applications, a defect-free polymer would be desirable to enhance th e conjugation in the backbone. X CH2SR R H2C S R R X OH X CH SR R H2C S R R X CH SR R H2C X CH SR R H2C X R2S CH SR R X n n Elim PPV SolublePrecursor Polymerization Figure 4-1. Indirect synthe sis of PPV via a soluble poly mer precursors established by Wessling Prior to the work presented in this chap ter, pure unsubstituted divinylbenzene was synthesized from terephthaldehyde via a Wittig reaction with potassium t -butoxide in THF. When divinylbenzene was added to the second-generation Grubbs catalyst ( 6 ) (500:1 monomer to catalyst ratio) in the presence of BHT (2.5 % by weight) at 50 C, the reaction mixture solidified within 5 min. Th e temperature was increased without melting the solid over the course of a week (>100 C at the end). Only the monomer and the dimer were ever observed by NMR in CDCl3 since the higher order species are presumably insoluble due to their intractabi lity and observation of an insoluble green solid in the solution even with heat. The amount that dissolve d in the NMR solution decreased as time progressed, and eventually no DVB or PPV peak could be seen in the
PAGE 120
103 spectra after 3 d at 100 C. Although molecular weight co uld not be directly measured here, this trend of decreasing solubility implied that molecular weight was increasing in the solid state. Since monomer and dimer we re observed in solid solutions, we concluded that the lower molecular weight species did indeed react in the solid state, as they were absent from the spectra in late r stages of the reaction. Howe ver, this could potentially be due to an increase in temperature above th e melting point of the lower melting species (monomer and dimer), giving these moieties the mobility they need to react and become part of an intractable polymer/oligomer domain. This prompted us to choose substitute d-divinylbenzene monomers to actually measure a molecular weight increase in the so lid state. Solublizing groups can be added to rigid-rod monomers to aid in macrom olecular characterization. Dialkoxy-PPVs are common in the literature because they can be processed, and exhibit interesting optical and electronic properties for use in light-emitting diodes184 and photovoltaic devices.185 We chose this class to more easily monitor ADMET reactions in the solid state, allowing further characterization in solution unattainable from its intractable parent polymer. Similar dialkoxy-divinylbenzene monomers to the one presented in this work, that contain a diheptoxy substitution pattern, have previously been shown by Thorn-Csnyi and coworkers186,187,188 to undergo ADMET, where they re port preparation of oligomers (mainly tetramer) in solution. Therein, they proposed that one of the oxygens from an alkoxy group forms a stable five-membered ring complex with the catalyst, which inhibits polymerization to an extent. A lthough SchrockÂ’s catalyst was mainly used in their work, a similar complex likely exists for Grubbs ruthenium catalysts ( 5 and 6 ); thus, it would be expected that only oligomers would form in the re sults reported herein. This
PAGE 121
104 work was designed to generate rigid-rod polymers operating under the conditions of solid-state ADMET demonstrated in Chapte r 3, in order to better understand the feasibility of increasing the molecular weight of rigid-rod polymer in the solid state, which could include the eventual synthesis of defect-free PPV. 4.2 Experimental 4.2.1 General All reactions were conducted in flame-dr ied glassware under an argon atmosphere unless otherwise noted. Reagents were purchased from Fische r or Aldrich and were used as received unless otherwise noted. Met hyl triphenyl phosphonium iodide was made from methyl iodide and triphenyl phosphine under standard Wittig conditions in ether. Metathesis catalysts used in this chapter, first( 5 )98 and second( 6 )103 generation Grubbs, were synthesized by known methods. Instrumentation for GPC and NMR are discussed in Chapter 2. Elemental analyses were performed by Atlantic Mi crolab, Inc (Norcross, GA). Mass spectrometry (MS) was conducted by the University of Florida Mass Spectrometry Services. Melting points were performed on a Laboratory Devices MelTemp apparatus and are not corrected. 4.2.2 Monomer Syntheses 4.2.2.1 Synthesis of 2,5-dibromohydroquinone (25) This procedure was modifi ed from the literature.189 Under atmospheric conditions, hydroquinone (100 g, 0.908 mol, 1.0 equiv) was a dded to an apparatus equipped with a potassium hydroxide (~0.02 M) trap to quench the HBr gas evolved from the reaction. Dichloromethane (DCM, 250 mL) and glacial acetic acid (250 mL) were added and stirred. Bromine (100 mL, 1.94 mol, 2.1 equi v) was added dropwise. As the first equivalent of bromine was added, the hydroquinone dissolved since the monobromo-
PAGE 122
105 intermediate is sol uble. The dibromo-product precipita ted from the orange solution upon the addition of the second equi valent. The reaction was allo wed to stir overnight. The precipitate was filtered and r ecrystallized in isopropanol to yield 121.3 g (50 %) of white 2,5-dibromohydroquinone ( 25 ) (MP = 196-202 C). 1H NMR (DMSO-d6, ppm): = 10.1-9.6 (br s, 2H), 7.04 (s, 2H). 13C NMR (DMSO-d6, ppm): = 147.29, 119.41, 108.21. HRMS: Calcd for C6H4Br2O2: 265.8578; Found: 265.8580. Elemental Anal. calcd for C6H4Br2O2: C 26.90, H 1.50, O 11.94, Br 59.65; Found: C 27.07, H 1.48, O 11.65, Br 59.39. 4.2.2.2 Synthesis of 1,4-dibromo-2,5-dipropoxybenzene (26) This compound was synthesized in a ma nner similar to that reported in the literature.189 2,5-Dibromohydroquinone ( 25 ) (70.0 g, 0.261 mol, 1.0 equiv) was dissolved in ethyl alcohol (200 mL). Potassiu m hydroxide (38.1 g, 0.679 mo l, 2.6 equiv) was dissolved in deionized water ( 300 mL), added dropwise, and allowed to stir for 1 h. A solution of 1-bromopropane (120 mL, 1.31 mo l, 5.0 equiv) in ethanol (100 mL) was prepared and added dropwise. The reaction fl ask was immersed in an oil bath at 50 C and allowed to stir with a mechanical stirre r overnight. The solution was dark orange with a white precipitate. The precipitate was filtered, washed with co ld, deionized water, and recrystallized with a 3:1 mixture of ethanol and benzene to yield 65.6 g (71 %) of white crystals of 1,4-di bromo-2,5-dipropoxybenzene ( 26 ) (MP = 79-81 C). 1H NMR (CDCl3, ppm): = 7.09 (s, 2H), 3.92 (t, 4H, J = 6.5 Hz), 1.83 (sex tet, 4H, J = 7.2 Hz), 1.06 (t, 6H, J = 7.2 Hz). 13C NMR (CDCl3, ppm): = 150.29, 118.74, 111.38, 71.99, 22.76, 10.73. HRMS: Calcd for C12H16Br2O2: 349.9517; Found: 349.9518. Elemental
PAGE 123
106 Anal. calcd for C12H16Br2O2: C 40.94, H 4.58, O 9.09, Br 45. 39; Found: C 41.13, H 4.61, O 9.20, Br 45.32. 4.2.2.3 Synthesis of 2,5-dipropoxyterephthalaldehyde (27) This procedure was modified from the literature.190 1,4-Dibromo-2,5dipropoxybenzene ( 26 ) (65.1 g, 0.185 mol, 1.0 equiv) was dissolved in dry diethyl ether (~900 mL) and cooled to 0 C. n -Butyl lithium (150 mL of 3.2 M, 0.461 mol, 2.5 equiv) was added dropwise over the course of an hour during which a white precipitate formed. N N -Dimethylformamide (DMF, 14 0 mL, 1.81 mol, 10 equiv) was added dropwise for an hour and the reaction was allowe d to stir overnigh t with a mechanical stirrer. The reaction was quenched with 5 % HCl (~500 mL ), turning the solution yellow. The organic layer was washed once with deionized water, and then the aqueous portion was extracted with three times DCM. The co mbined organic portions were dried over magnesium sulfate. Filtration and solvent evaporation yielded crude product, which was purified by column chromatography on silica gel with a 3:1 mixture of toluene and hexanes, followed by a recrystallizati on in hexanes to yield 10.6 g (23 %) of 2,5dipropoxyterephthalaldehyde ( 27 ) (MP = 80-82 C, lit190 = 75-76 C). 1H NMR (CDCl3, ppm): = 10.53 (s, 2H), 7.44 (s, 2H), 4.06 (t, 4H, J = 6.6 Hz), 1.87 (sextet, 4H, J = 7.2 Hz), 1.07 (t, 6H, J = 7.2 Hz). 13C NMR (CDCl3, ppm): = 189.66, 155.44, 129.50, 111.85, 70.90, 22.66, 10.73. HRMS: Calcd for C14H18O4: 250.1205; Found: 250.1208. Elemental Anal. calcd for C14H18O4: C 67.18, H 7.25, O 25.57; Found: C 67.23, H 7.25, O 25.43. 4.2.2.4 Synthesis of 1,4-diprop oxy-2,5-divinylbenzene (28) 2,5-Dipropoxyterephthalaldehyde ( 27 ) (3.20 g, 0.0128 mol, 1.0 equiv) and methyl triphenyl phosphonium iodide (12.92 g, 0.0320 mol, 2.5 equiv) were dissolved in dry
PAGE 124
107 THF (~200 mL), which was transferred via a cannula. A 1.0 M solution of potassium t -butoxide in THF (55 mL, 0.055 mol, 4 equiv) was added dropwise over a half an hour, during which the solution changed from yello w to brown. The flask was place in an oil bath at 70 C and was allowed to stir overn ight. Deionized water (~100 mL) was added to quench the solution. Diethyl ether (~200 mL) was added and the organic portion was washed with deionized water, while the a queous portions were ex tracted twice with diethyl ether. The combined organic layers were washed with brine, and then dried over magnesium sulfate. Filtration and solvent evaporation yielded 11.08 g of crude material, which was flashed through a plug of silica, and then purified by column chromatography on silica gel with a solvent mixture of 10 % benzene in hexanes to yield 1.63 g (52 %) of white 1,4-dipropoxy-2,5divinylbenzene ( 28 ) (MP = 78-80 C) crystals. 1H NMR (CDCl3, ppm): = 7.05 (dd, 2H, J = 18.0 Hz, 11.1 Hz), 6.99 (s, 2H), 5.73 (dd, 2H, J = 18.0 Hz, 1.5 Hz), 5.23 (dd, 2H, J = 11.1 Hz, 1.5 Hz), 3.93 (t, 4H, 6.6 Hz), 1.83 (sextet, 4H, 7.2 Hz), 1.05 (t, 6H, 7.2 Hz). 13C NMR (CDCl3, ppm): = 150.86, 131.76, 114.24, 110.81, 71.09, 23.01, 10.91. HRMS: Calcd for C16H22O2: 246.1620; Found: 246.1610. Elemental Anal. calcd for C16H22O2: C 78.01, H 9.00, O 12.99; Found: C 78.01, H 9.01, O 12.87. 4.2.2.5 Synthesis of 1,4-dibromo-2,5-dioctadecyloxybenzene (29) This compound was synthesized in a manne r analogous to that reported in the literature.189 2,5-Dibromohydroquinone ( 25 ) (17.9 g, 0.0668 mol, 1.0 equiv) was dissolved in THF (200 mL) in a 2 L threeneck round-bottom flas k equipped with an addition funnel and condenser. Potassium hydroxide (9.76 g, 0.174 mol, 2.6 equiv) was dissolved in ethyl alcohol (150 mL), added dropwise, and allowed to stir for 1 h. A solution of 1-bromooctadecane (49.0 g, 0.147 mol, 2.2 equiv) in THF (500 mL) was
PAGE 125
108 prepared and added dropwise during which the so lution turned to dark orange and then to brown. The reaction flask was imme rsed in an oil bath at 90 C and allowed to reflux for three days. During the course of the reacti on a white precipitate fo rmed from the brown solution. Upon cooling to room temperatur e, deionized water (400 mL) was added to encourage more precipitation. The precipitate was filtered, washed with cold, deionized water, and recrystallized three times in chloroform to yield 30.5 g (59 %) of white crystals of 1,4-dibromo2,5-dioctadecyloxybenzene ( 29 ) (MP = 90-92 C). 1H NMR (CDCl3, ppm): = 7.08 (s, 2H), 3.94 (t, 4H, J = 6.6 Hz), 1.83 (p, 4H, J = 6.9 Hz), 1.48 (m, 4H) 1.42-1.16 (br, 56 H), 0.88 (t, 6H, J = 6.6 Hz). 13C NMR (CDCl3, ppm): = 150.31, 118.70, 111.36, 70.54, 32.16, 29.93, 29.89, 29.81, 29.78, 29.59, 29.52, 29.34, 26.15, 22.92, 14.35. HRMS: Calcd for C42H76Br2O2: 770.4212; Found: 770.4211. Elemental Anal. calcd for C42H76Br2O2: C 65.27, H 9.91, O 4.14, Br 20.68; Found: C 65.17, H 9.96, O 4.14, Br 20.41. 4.2.2.6 Synthesis of 1,4-dioctadecy loxy-2,5-divinylbenzene (30) This compound was synthesized in a manne r analogous to that reported in a procedure from the Reynolds group.191 Toluene was distilled from sodium metal and stored in a Schlenk flask before use. 1,4-Dibromo-2,5-dioc tadecyloxybenzene ( 29 ) (5.00 g, 0.00647 mol, 1 equiv) was added to a 500 mL three-neck roundbottom flask equipped with a condenser, a stopper, and a septa with positive argon pressure. To this flask, dry toluene (150 mL) was added via cannula, a nd then tributyl-vinyl-stannane (4.51 g, 0.0142 mol, 2.2 equiv) was added via syringe. After increasing the temperature to 80 C to dissolve all contents, the resultant solu tion was degassed by bubbling argon through a needle for 25 min, after which tetrakis (triphenylphosphine) palladium(0) (Pd(PPh3)4)
PAGE 126
109 (0.299 g, 0.000259 mol, 0.04 equiv) was added with positive argon pressure. The solution, initially yellow, was refluxed at 130 C for 2 d, changing to a blue within the first few hours. Upon cooling, the solution was added to deionized water (150 mL) and extracted twice with diethyl ether (200 mL, 50 mL). The co mbined organic layers were washed with brine and dried over MgSO4. Upon filtration and so lvent evaporation, the crude product was purified by column chroma tography using 10 % benzene in hexanes to yield 1.23 g (29 %) of white 1,4-dibromo-2,5-dioctadecyloxybenzene ( 30 ) (MP = 78-79 C). 1H NMR (CDCl3, ppm): = 7.04 (dd, 2H, J = 18.0 Hz, 11.1 Hz), 6.98 (s, 2H), 5.72 (dd, 2H, J = 18.0 Hz, 1.5 Hz), 5.24 (dd, 2H, J = 11.1 Hz, 1.5 Hz), 3.96 (t, 4H, 6.3 Hz), 1.80 (p, 4H, 7.2 Hz), 1.47 (m, 4H), 1.41-1.20 (br, 56 H), 0.88 (t, 6H, 6.6 Hz). 13C NMR (CDCl3, ppm): = 150.88, 131.80, 127.41, 114.21, 110.78, 69.57, 32.16, 29.94, 29.85, 29.83, 29.69, 29.65, 29.60, 26.38, 22.93, 14.35. HRMS: Calcd for C46H82O2: 666.6315; Found: 666.6321. Elemental Anal. calcd for C46H82O2: C 82.81, H 12.39, O 4.80; Found: C 82.62, H 12.51, O 4.99. 4.2.3 Solid-State Polymerization Conditions Methods of creating a solid solution, runni ng a solid-state reaction, modes of quenching, and types of molecula r weight characterization are discussed in Chapter 2. In addition, BHT (0.1 to 5.0 wt %) was also added to the mixtures of monomer and catalyst in order to help prevent radical polymeriza tion during the solid-state reaction. This was done since vinylbenzene-type olefins are more susceptible to radical chemistry than those in Chapter 3. However, it is unknown if BHT has the same radical scavenging ability in the solid state. The details of what was us ed for each particular experiment including catalyst loading, the amount of BHT, and reaction temperature are specified in the results
PAGE 127
110 and discussion section as each experiment is conveyed. In a typical experiment, each sample was analyzed for molecular weight by GPC and/or NMR end-group analysis. Any additional characterization will appear in the results and discussion as well. Full characterization has not been included since on ly trends in molecular weight increase in the solid state were desired. 4.3 Results and Discussion Two dialkoxy-divinylbenzene monomers were synthesized for this work. The first is the dipropoxy, show n in Figure 4-2.192 Hydroquinone was dibr ominated to yield 2,5dibromohydroquinone ( 25 ) following the literature.189 Compound 25 was then alkylated using Williamson ether synthesis to yield 1,4-dibromo-2,5-dipropoxybenzene ( 26 ).189 This was then acetylated using butyl lithium and N N -dimethylformamide to form the dialdehylde 27 .190 Subsequent reaction using standard Wittig conditions yields 1,4dipropoxy-2,5-divinylbenzene ( 28 ), which can be used in an ADMET polymerization. O O OH HO OH HO Br BrBr O O Br O O O O H H Br2CH2Cl2Acetic Acid Br KOH / EtOH CH3PPh3I t -BuOK THF 2526 2728 DMF ether n -BuLi Figure 4-2. Synthetic route to 1,4-dipropoxy-2,5-divinylbenzene ( 28 ) Even with the two propoxy groups on this monomer, the polymer still has some solubility issues. The second monomer synt hesized had much longer alkoxy chains with
PAGE 128
111 18 carbons. This synthesis of this mono mer, 1,4-dioctadecyloxy-2,5-divinylbenzene ( 30 ), is shown in Figure 4-3. As in th e previous scheme, 2,5-dibromohydroquinone ( 25 ) was alkylated using Williamson ether synthesi s, this time with the 18-carbon chain bromide to yield the 1,4-dibro mo-2,5-dioctadecyloxybenzene ( 29 ).189 A subsequent Stille coupling provides route direc tly to the divinyl product ( 30 ) from the dibromide.191 This is in contrast to the synthesis of 28 in Figure 4-2, which did th is conversion in two steps from 26 rather than one. SnBu3Br O O Br 17 17 O O 17 17 OH HO Br Br 2930 25 Br 17 KOH / EtOHPd(PPh3)4 Figure 4-3. Synthetic route to 1,4 -dioctadecyloxy-2,5-divinylbenzene ( 30 ) The solid-state conversion of an oligomer of monomer 28 via second-generation Grubbs catalysis ( 6 ) is shown in Figure 4-4. 1,4-Dipropoxy-2,5-divinylbenzene ( 28 ) (0.50 g, 0.0020 mol) was added to the typical solid-state apparatus. Approximately 0.2 wt % of BHT (0.0010 g) was added to prevent any radical polymerization. The reaction flask was brought to 80 C to melt the monomer, at which point secondgeneration Grubbs catalyst ( 6 ) (5.7 mg, 0.0067 mmol) was adde d in a 300:1 monomer to catalyst ratio and allowed to react for 5 min. The reaction mixt ure solidified upon cooling to room temperature, and polymerizat ion was continued in the solid state at 45 C for 1 week. This polymerization was carried out in th e same manner as shown in Figure 3-3 for 1,9-decadiene ( 11 ), whereby the reaction was started in the melt, and then continued in the solid state. Since divinyl benzene molecules are prone to radical chemistry that would
PAGE 129
112 lead to cross-linking through the vinyl groups, a free radical inhibitor, 2,6-dit -butyl-4methylphenol (BHT), was ground in to the reaction mixture to i nhibit this chemistry. This particular experiment was conducted w ith a slightly impure batch of compound 28 which may be a factor in this work. As expecte d, only oligomer was formed in both the melt and solid-state reactions. The significant findi ng here is that the molecular weight has increased while in the solid state. The num ber average molecular weight was effectively doubled in the solid state from the 500 g/ mol achieved in the initial melt reaction. 6 BHT 5 min at 80 oC (melt) 6 BHT 7 Days at 45 oC (solid state) Mn = 500 g/mol Mw = 900 g/mol Mn = 1100 g/mol Mw = 2100 g/mol O O 28 O O x O O y Figure 4-4. Prepolymeri zation in the melt of 28 followed by polymerization in the solid state using second-generation Grubbs ruthenium catalyst ( 6 ) (Number average molecular weights ( Mn) and weight average molecular weights ( Mw) were obtained by GPC in THF relative to polystyrene standards.) Similar experiments were run with th e first-generation Grubbs catalyst ( 5 ) (300:1 monomer to catalyst ratio). One particul ar run was started in the melt at ~100 C for 8 min with 0.5 wt % BHT with subsequent increase from 300 g/mol to 600 g/mol in 4 d at 60 C in the solid state by GPC in THF relative to polystyrene standards. Less significant of an increase was expected since the first-generation catalyst decomposes in the melt portion of the experiment as seen in Chapter 3. In addition, it is hard to draw any concrete conclusions with these low numbe rs since they are within the error of the technique. The second monomer with the longer alkoxy ch ains only saw very slight increases in molecular weight in the solid state. In one experiment with second-generation Grubbs
PAGE 130
113 ( 6 ) (300:1 monomer to catalyst ratio) and 1.0 wt % BHT, afte r a melt period of 5 min at 120 C, an increase from 2,200 g/mol to only 2,400 g/mol was observed in the solid state after 7 d at 60 C by GPC in THF relative to polystyrene. This almost negligible increase may be an indication of the difficulty that rigid-rod polymers may have in undergoing this type of polymerization. However, any in crease is significant since these polymers do not possess the same freedom of molecular mo tion that the semicrystalline polymers in Chapter 3 possess. 4.4 Conclusions The feasibility of solid-state ADMET of rigid-rod polymers was investigated with two dialkoxy-divinylbenzene monomer. In one experiment with the dipropoxydivinylbenzene monomer ( 28 ), an effective doubling of molecular weight was observed in a similar experiment to the flexible polymer in Chapter 3. In other experiments with monomers 28 and 30 only slight increases in molecu lar weight were observed in the solid state. However, it is important to note that any increase of molecular weight using a rigid-rod polymer is significant since they are operating under the same principles as flexible polymer chain.
PAGE 131
114 CHAPTER 5 SOLID-STATE RING-CLO SING METATHESIS 5.1 Introduction One of the most intriguing aspects of olef in metathesis is that several different types of chemistry can be performed dependi ng on reaction conditions and the structural features of the substrate w ith the same carbene catalysts Ring-closing metathesis (RCM)50,51,52,53 is widely used in organic synthesis to obtain cyclic structures containing a carbon-carbon double bond. In addition to being found in an uncountable number of natural products, the double bond offers functiona lity that can be subsequently used in a number of other chemistries, providing anot her tool to the organi c chemist in making chemical conversions. This work is the first aimed at conducting metathesis condensation reactions in the solid state to generate small molecules rather than macromolecules. Unlike ring opening metathesis polymerization (ROMP),42,56,57,58 which is a chain addition reaction, RCM is an equilibrium reaction where ethylene is evolved as a condensate in the same manner as ADMET. Fi gure 5-1 illustrates the similarity of the ADMET and RCM mechanisms. The only differenc e is that RCM is an intramolecular reaction that leads to ring-closed products, while ADMET is an intermolecular reaction that leads to polymer. Ring-closing meta thesis and ADMET are often in competition with one another, where ring size is the determ inate factor as to which primarily occurs. Since ADMET works in the solid state by the removal of the volatile condensate, ethylene, there is potential that RCM can undergo a similar conversion.
PAGE 132
115 R n -1 R R R RuCH2R Ru ethylene R Ru R Ru R R RuCH2R Ru ethylene R Ru R R RuADMET Mechanism RCM Mechanismring-closed product polymer Figure 5-1. Comparison of th e ADMET and RCM mechanisms An exploration into the feasibility of conducting RCM in the so lid state could be invaluable to the organic community. Dis tinct advantages would exist when conducting this reaction in the solid state. Lower reaction temperatures reduce or eliminate isomerization of double bonds and other side re actions, which could lead to higher yields and higher product purity. Many organic target molecules, as well as intermediates along a synthetic route could have solubility issu es. If this conversion could be accomplished in the solid state, this probl em could be easily circumvented by performing the RCM step in the solid state. In addition to being potentially valuable to the organic discipline, this work may result in the unveiling of c oncepts that are present in ADMET polymerization since these reactions share the same mechanistic princi ples. The diene star ting materials would presumably be crystalline in nature. Since it is generally accepted that solid-state polycondensations occur in the am orphous region of semicrysta lline polymers, this could mean that reaction in the solid state is unlikel y. However, the same could be said for the monomer-to-polymer experiments in Chapter 3 where conversions to oligomer were observed in the solid state. The fact that any conversion occurs in an environment such as this is significant.
PAGE 133
116 Several small dienes are commonly used in the literature to test catalyst activity towards RCM.103, 193,194 A handful of these are shown in Fi gure 5-2. It is desirable to use these same dienes when testing a particular aspect of the way RCM is conducted such as in this work. However, these common dienes could not be used in this solid-state study because of the simple fact that all of these common dienes are liquids at room temperature. New dienes that were solids at room temperature were synthesized for this study. EE N O O Ts OH OTBS Ph E E Figure 5-2. Common RCM dienes found in the literature for testing features of RCM (E = CO2Et) Herein, the solid-state RCM activity was tested with four solid dienes and six catalyst systems. Multiple modes of creating a solid solution were tested including the grinding of solid diene and catalyst togeth er and freeze-drying. Recently, reports have shown that microwave-induced liquid RCM is possible,194,195,196,197,198,199 the advantage being shorter reaction times.200 Consequently, reactions unde rgo fewer side reactions and have a more simplistic work-up procedure to yield cleaner products in higher yields. Microwave RCM reactions can be performed with solvent or without solvent (bulk state).195 The effects of microwaves on solid -state RCM were explored herein.
PAGE 134
117 5.2 Experimental 5.2.1 General All reactions were conducted in flame-dr ied glassware under an argon atmosphere unless otherwise noted. Reagents were purchased from Fische r or Aldrich and were used as received unless otherwise noted. Metathesis catalysts used in this chapter, first( 5 )98 and second( 6 )103 generation Grubbs, Sc hrockÂ’s molybdenum ( 2 ), 83 HoveydaÂ’s catalyst ( 7 ),109 MolÂ’s catalyst ( 8 ),110,111 and the Hoveyda-Mol hybrid ( 9 )112 were synthesized by known methods. Instrumentation for GPC and NMR are discussed in Chapter 2. Elemental analyses were performed by Atlantic Microl ab, Inc (Norcross, GA). Mass spectrometry (MS) was conducted by the University of Fl orida Mass Spectrometry Services. Melting points were performed on a Laboratory De vices Mel-Temp appa ratus. Microwave experiments were run on a single-mode CEM Discover Microwave Synthesizer producing continuous irra diation at 2455 MHz. 5.2.2 Synthesis of Dienes for Solid-State RCM 5.2.2.1 Synthesis of N N -diallyl-3,5-dinitrobenzamide (31) Triethylamine and DCM were dried over f our-angstrom molecular sieves prior to use. 3,5-Dinitrobenzoyl chloride (5.0 g, 0.217 mol, 1.0 equiv) were added to a Schlenk flask, to which dried DCM was added via a cannula. Dried triethylamine (4.0 mL, 1.3 equiv) and diallylamine (3.0 mL, 0.0239 mol, 1.1 equiv) were added to the reaction flask via a syringe. With DCM still adding to the Schlenk flask, the acyl chloride solution was transferred to the reaction flask via cannula, at which point the solution turned brown. With approximately half the acyl chloride solution added, the remainder was poured directly into the reaction flask. The soluti on was allowed to stir overnight, during which
PAGE 135
118 time the brown color persisted. The solution wa s washed with 3 N HCl, saturated sodium bicarbonate, and brine, and then dried over magnesium sulfate. Filtration and solvent evaporation yielded 5.15 g (82 %) of crude product that was r ecrystallized in ethanol with slow cooling to yield light yellow crystals of N N -diallyl-3,5-dinitrobenzamide ( 31 ) (MP = 65-67 C). 1H NMR (CDCl3, ppm): = 9.09 (t, 1H, J = 2.1 Hz), 8.66 (d, 2H, J = 2.1 Hz), 6.0-5.7 (br s, 2H), 5.45-5.20 (m, 4H), 4.20 (br s, 2H), 3.85 (br s, 2H). 13C NMR (CDCl3, ppm): = 166.95, 148.58, 139.57, 132.13, 131.83, 127.32, 119.94, 119.37, 118.74, 50.90, 48.26. HRMS: Calcd for C13H14N3O5 ([M+H]+): 292.0933; Found: 292.0932. Elemental Anal. calcd for C13H13N3O5: C 53.61, H 4.50, N 14.43, O 27.47; Found: C 53.70, H 4.51, N 14.42, O 27.55. 5.2.2.2 Synthesis of 1-allyl-3-but enyl octadecanoate (32) Stearic acid (13.2 g, 0.0464 mol, 1,3 equiv) was dissolved in dichloromethane (150 mL) with heat. Diisopropyl car obodiimide (DIC, 8.4 mL, 6.75 g, 0.0535 mol, 1.5 equiv) was added, turning the so lution a turbid yellow. The solution was allowed to stir for 2 h before the 1,6-hepta dien-4-ol (4.0 g, 4.63 mL, 0.035 7 mol, 1 equiv) was added along with a spatula tip of 4-dimethylaminopyr idine (DMAP), and allowed to reflux at 45 C overnight. Upon cooling, the solution wa s washed twice with saturated sodium bicarbonate, once with brine, and dried over magnesium sulfate. Filtration and solvent evaporation yielded 20.29 g of crude prod uct, which was purified with column chromatography with 5 % ethyl acetate in hexanes to yiel d 9.17 g (68 %) of the white 1-allyl-3-butenyl octadecanoate ( 32 ) (MP = 28-30 C) crystals. 1H NMR (CDCl3, ppm): = 5.83-5.67 (m, 2H), 5.13-5.03 (m, 4H), 4.98 (p, 1H, J = 6.3 Hz), 2.40-2.23 (m, 4H), 2.27 (t, 2H, J = 7.5 Hz), 1.60 (p, 2H, J = 7.5 Hz ), 1.36-1.18 (br, 28H), 0.88 (t, 3H, J = 6.6 Hz). 13C NMR (CDCl3, ppm): = 173.59, 133.83, 117.99, 72.20, 38.28, 34.80, 32.15,
PAGE 136
119 29.91, 29.87, 29.83, 29.70, 29.59, 29.49, 29.35, 25.29, 22.91, 14.33. HRMS: Calcd for C25H46O2: 378.3498; Found: 378.3490. Elemental Anal. calcd for C25H46O2: C 79.30, H 12.25, O 8.45; Found: C 79.29, H 12.35, O 8.55. 5.2.2.3 Synthesis of diallyl malonic acid (33) Potassium hydroxide (7.50 g, 0.134 mol) was dissolved in ethanol (50 mL). Diethyl diallylmalonate (5.00 g, 0.208 mol) was added and allowed refluxed overnight at 70 C. After cooling, the ethanol was eva porated, deionized wate r (~100 mL) was added, and the solution was acidified with concen trated HCl. The diacid product was then extracted three times with ether, washed w ith brine, and then dried over magnesium sulfate. Filtration and solvent evaporati on yielded 2.26 g (59.0 %) of the crude product, which was recrystallized from hexanes/ether w ith slow evaporation of the ether to yield 1.54 g (69 % recovery, 40 % over all) of fine, white needle s of diallyl malonic acid ( 33 ) (MP = 137-142 C). The crystals can be decolo rized with decolorizing carbon in ether if they are not white. 1H NMR (CDCl3, ppm): = 9.7-8.4 (br, 2H), 5.80-5.64 (m, 2H), 5.23-5.13 (m, 4H), 2.69 (d, 4H, J = 7.5 Hz). 13C NMR (CDCl3, ppm): = 176.28, 131.60, 120.31, 58.09, 37.85. ESI-MS: Calcd for C9H11O4 ([M-H]-): 183.1; Found: 183.1. Elemental Anal. calcd for C9H12O4: C 58.69, H 6.57, O 34.75; Found: C 58.70, H 6.61, O 34.82. 5.2.2.4 Synthesis of the dipotassium salt of diallyl malonic acid (34) Diallyl malonic acid ( 33 ) (0.600 g, 0.00326 mol) was dissolved in ethanol (~40 mL). A solution of potassium hydroxi de (1.094 g, 0.0195 mol) in ethanol (~40 mL) was added dropwise as a white precipitate slowly formed from solution. The milky, white solution was allowed to stir for a half an hour and then cooled in an ice bath. The white, pasty precipitate was f iltered, washed with cold ethanol, and dried under vacuum
PAGE 137
120 to yield 0.630 g (74 %) of th e white dipotassium salt ( 34 ) (MP = > 250 C). 1H NMR (D2O): = 5.80-5.62 (m, 2H), 5.17-5.02 (m, 4H), 2.46 (d, 4H, J = 7.2 Hz). 13C NMR (D2O): = 181.0, 135.5, 117.5, 61.6, 37.7. ESI-MS: Calcd for C9H11K2O4 ([M+H]+): 261.0; Found: 261.0. Elemental Anal. calcd for C9H10O4K2: C 41.52, H 3.87, O 24.58, K 30.03; Found: C 41.32, H 3.99, O 24.70. 5.2.3 Solid-State RCM Reactions Conditions Methods of creating a solid solution, runni ng a solid-state reaction, modes of quenching, and types of molecula r weight characterization are discussed in Chapter 2. A typical solid-state RCM reaction was performed with a 300:1 monomer to catalyst ratio scaled to 0.40 g of monomer, although smaller scales were occasionally used. The solid monomer and catalyst were mixed and ground toge ther in a mortar and pestle three times in atmospheric conditions, except for when Schrock’s catalyst was used, which was ground in a glovebox under argon. The solid mixture was added to a reaction vessel with argon flow to a bubbler and immersed in a temperature controlled oil bath. The reaction progress was monitored periodically by NMR, in cluding a “start sample” from each run prior to the temperature increase. In all cas es, this starting sample showed no reaction. A typical experiment was run for 1 week. At the end of each experiment, an unquenched portion was allowed to react in solution, us ually the NMR solvent, to show that the catalyst was still active. The catalyst was found to still be active in all cases unless otherwise noted. Percent yield for diene 31 were calculated by comparing the integrations of the aromatic protons ortho to the amide car bonyl (8.66 ppm) to those of the RCM product ( 35 ) (8.76 ppm). The experiment with diene 32 was conducted at reduced temperatures (20 C) due to it low melting point. Diene 33 was found to react with EVE (a notable by-
PAGE 138
121 product was acetylaldehyde from the degradation of EVE) and could not be quenched in the standard manner. Instead, a sample was allowed to sit under atmospheric conditions for approximately 3 weeks after the reactionÂ’s conclusion in order to minimize catalyst activity in the NMR solution. 5.2.4 Synthesis of Ring-Closed Product (2,5dihydropyrrol-1-yl)-(3,5-dinitrophenyl)methanone (35) The ring-closed product ( 35 ) of diene 31 was synthesized in solution to know what peaks to look for in the NMR during the so lid-state polymerizati on. Although careful purification was not performed, NMR and HRMS confirmed the presence of the product. To obtain an elementally pure sample, fu rther purification should be performed. Chloroform was added to a solid-stat e reaction containing 0.40 g of diene 31 and 3.8 mg of catalyst 5 after a 7-d reaction time and then allowe d to stir overnight. The solvent was then evaporated and the crude product was recr ystallized in ethanol to yield 0.31 g (86 %) of light brown needles of (2,5-dihydropy rrol-1-yl)-(3,5-dinitr ophenyl)-methanone ( 35 ) (MP = 156-157 C). 1H NMR (CDCl3, ppm): = 9.12 (t, 1H, J = 2.1 Hz), 8.76 (d, 2H, J = 2.1 Hz), 6.14-5.96 (m, 1H), 5.86-5.78 (m, 1H ), 4.56-4,48 (m, 2H), 4.34-4,26 (m, 2H). 13C NMR (CDCl3, ppm): = 164.78, 148.70, 140.02, 127. 61, 126.40, 124.82, 120.13, 55.95, 54.27. HRMS: Calcd for C11H9N3O5: 263.0542; Found: 263.0546. 5.3 Results and Discussion Ring-closing metathesis shares the same equilibrium-type metathesis mechanism as ADMET, where the conversion is dependent on the removal of the condensate, ethylene. It is for this similarity to ADMET polymeri zation that we began our investigation into performing RCM in the solid state. Several small dienes are commonly used to test catalyst activity towards RCM,103 ,193,194 all of which are liquids at room temperature.
PAGE 139
122 Consequently, new dienes (Figure 5-3), whic h are solids at room temperature, were synthesized to in order to attempt solid-state RCM. N OO OH HO OO OK KO O NO2NO2O n -C17H36O 31 32 34 33 Figure 5-3. Diene synthesized fo r solid-state RCM experiments The diene primarily used in this study was amide 31 which had a melting point of 67-69 C; thus, solid-state RCM reactions could be conducted at temperatures that we have shown to be active with ADMET chemistry in the solid state (35-45 C).112 Long chain ester diene 32 was synthesized to yield a solid with a melting point very close to room temperature (MP = 28-30 C). Experiments with this di ene in the solid state could only be conducted at reduced temperatures. The diacid 33 and dicarboxylate 34 were also synthesized with functiona lity that would render it a solid at room temperature. However, the acid and carboxylate functionality on these dienes may have interactions with the ruthenium metathesis catalysts, inhibiting the metathesis mechanism. Six metathesis catalysts ( 2, 5-9 ) were employed in this re search (shown in Figures 1-17 and 1-19). Experiments were conducted with each using amide 31 as the diene. The diene and the catalyst were mixed and gr ound together (300:1 ratio) in a mortar and pestle three times, resulting in a mixture with a melting point only 2 lower than pure diene. The solid mixture was then placed in the reaction vessel at 35 C (room
PAGE 140
123 temperature for catalyst 2 ) with argon flow and monitore d for 1 week. The results are shown in Table 5-1 (lines 1-6). Table 5-1. Results of solid-state RCM experiments Line Diene Cat. Temp ( C) Time (d) Yield 1 31 5 35 7 2.3 % 2 31 6 35 7 NR 3 31 2 25 7 NR 4 31 7 35 7 NR 5 31 8 35 7 0.6 % 6 31 9 35 7 1.8 % 7 31 5 45 1 3.2 % 8 32 8 20 7 NR 9 33 6 45 7a > 1.1 %b 10 34 5 60 7 NR 11 34 5 100 7 NR 12 34 5 140 7 NR a After the reaction period, the sample sat in atmospheric conditions for approximately 3 weeks before sampling. b The 1.1 % yield is from an unquenched sample and any to all of this could be from reaction in solution. Solid-state RCM was observed using catalysts 5 8 and 9 albeit in very low yields, while no apparent reaction was observed using 2 6 and 7 Although the yields with 5 8 and 9 are indeed low, it is significant that an y conversion has been achieved here because it is occurring in the solid state. Of the three catalysts that promoted RCM under these conditions, the first-gene ration Grubbs catalyst ( 5 ) appeared to be the most effective, reaching 2.3 % in a week; thus it was used in further experimentation with diene 31 This reaction scheme to yield this ring-closed product ( 35 ) is shown in Figure 5-4. Yields will undoubtedly be higher in solution as was demonstrated by the addition of chloroform to a solid mixture of 31 and 5 to achieve a high yield (86 %) of the ring-closed product 35 which was used to identify the peaks to look for in the NMR spectra. An experiment with 31 and 5 at 45 C is also shown in Table 5-1 (line 7), showing a slightly higher yield achieved (3.2 %) af ter only 1 d in the so lid state. Higher
PAGE 141
124 temperatures (50-55 C) lead to partial liquid systems resulting in higher yields, particularly within the liquid por tions. In an e xperiment at 45 C, the oil bath temperature increased (~50 C) for approximately 20 min duri ng the first day of reaction. A visibly heterogeneous darkening of the sa mple was observed for portions of the solid mixture. Careful sampling of the different ar eas of the mixture showed that the darkened portion had a higher yield (10.5 %) compared to the undarkened portion (4.9 %). Another experiment was perf ormed partially in the melt at 65-70 C for approximately 6 min (yield = 26 %), and then cooled to a solid over the next 13 min, at which point the yield was 27.4 %. The reaction was then placed at 45 C and monitored for 1 week. The percent yield was found not to change significan tly during the week (28.1 % at 2.5 h, 25.5 % at 2 d, and 27.2 % at 6 d) with some fluctuation in yield indicating a heterogeneous melt period. N O NO2NO2N O NO2NO25 31 35 + Figure 5-4. Ring-closing me tathesis of amide diene 31 to compound 35 using catalyst 5 The heterogeneous nature of these experime nts suggest that the crystal is collapsing around sites of defects, either from an actual ph ysical defect in the crystal or at places where an impurity exists, such as the catalyst, the RCM pr oduct, or possibly oligomer. This may be creating local non-crystalline environments where the reaction can occur without difficulty. As the reaction proceeds in the solid state, the presence of the product depresses local melting points, and at some point, incorporation of the product becomes
PAGE 142
125 impossible and deterioration of the crystal occurs. This can be visibly observed by a heterogeneous darkening of portions of the mixt ure. By careful sampling of the darkened and undarkened portions, it was determined that this change does occur between 4-10 % yield. This could be an indication that some amount of amorphous characte r needs to be present in order for this reaction to occur. Perhaps the reaction st arts by nucleation about sites of defects in the crystal, potentially from physical defects caused by the grinding process or the presence of catalyst in the mi xture. Then, growth about these centers of defects could propagate from the appearan ce of the RCM product in the mixture. Experiments with other dienes met w ith less success (Table 5-1). Diene 32 had a melting point too close to room temperature (28-30 C) to be performed at typical metathesis temperatures. However, Mol’s catalyst ( 8 ) has been shown to be active at reduced temperatures.112 An experiment with catalyst 8 at 20 C was performed (line 8) with no apparent reaction occu rring after 1 week. To combin e the two reactants in this experiment, grinding was done in a cooled mortar from being in an o dichlorobenzene/liquid nitrogen slurry (–15 C) with a funnel flowing nitrogen gas above to prevent condensation of wa ter vapor from entering the r eaction. The mixture was then placed in an ice bath (0 C) and allowed to warm to 20 C by the slow melting of the ice. The resultant solid mixture was sampled periodically for a week with no observed reaction. Diene 33 has carboxylic acid functionality, which unfortunately reacts with the quenching agent, ethyl vinyl ether. Th e mechanism of deco mposition is unknown, however, one notable side product was acetylaldehyde, which comes from the degradation of EVE. In the experiment sh own in Table 1-5 (line 9), the solid mixture
PAGE 143
126 was placed in a flask at 45 C and allowed react for 1 week. Since the typical quench procedure could not be used, the sample wa s instead allowed to sit under atmospheric conditions for approximately 3 weeks to aid in catalyst decomposition for minimal catalyst activity in the characterization so lution. An unquenched NMR sample was prepared and immediately analy zed to show a 1.1 % yield. None to all of this could have been from reaction in the NMR solution since complex 5 is quite durable in atmospheric conditions and may have still been somewhat active. Regardless, th e yield is quite low and did not warrant further study. For diene 34 chloroform-d could not be used as the NMR solvent since it has carboxylate functionality, a nd is therefore too pola r to dissolve in CDCl3. Deuterated water was used instead, which also doubled as the quenching agent since water degrades the catalyst as well. Three expe riments were conducted at 60, 100, and 140 C (lines 10-12) and are shown in Table 5-1. All three of these yielded no apparent reaction. This could however, be due to the carboxylate functionality somehow complexing to the ruthenium, and forming a species inactive towa rds metathesis with the catalyst. If the ruthenium were to react with one side of diene 34 the oxygen atoms on the carboxylates could form a six-membered ring by complexing to the ruthenium, so this is a feasible explanation for not seeing any RCM. Other methods of reaction and sample prep aration were explored using amide diene 31 and the first-genera tion Grubbs catalyst ( 5 ). The freeze-drying of diene and catalyst in benzene produced a solid mixture containing 1. 0 % yield of product. A trace amount of benzene remained as well (~150:1 diene to be nzene ratio by NMR). After 3 d of being exposed to vacuum at 35 C, the reaction yield increased to 4.6 % with similar amount of
PAGE 144
127 benzene remaining. This is slightly higher than the 2.3 % yield on line 1 of Table 5-1 conducted at the same temperat ure. However, 1.0 % came from reaction in a solution state, and the trace amount of benzene may have played a ro le in assisting the reaction. Recently reports have shown that microwave-induced liquid RCM is possible.194,195,196,197,198,199 The effect on microwaves to enhance the rate of RCM in the solid state was investigated for diene 31 and first-generation Grubbs ( 5 ). In 20 min of microwave exposure at 35 C, 0.7 % yield was achieved, wh ich is what was seen under thermal conditions after 24 h. This shows th at RCM reaction rates can be increased by microwave radiation in the solid state in addition to liquid states. Multiple microwave experiments were conducted in 5-h intervals at 55 C on the same sample to see how high the yield coul d be pushed without pa rtial melting of the solid solution. The heterogeneous darkeni ng of the sample was observed during the second 5-h period. This appeared to be the same as was observed in the thermal experiments described above. By careful sampling of th e darkened and undarkened portions, it was determined that this cha nge occurred between 410 % yield, which was what was observed thermally as well. 5.4 Conclusions Solid-state RCM to low conversions was demonstrated with amide diene 31 The first-generation Grubbs catalyst ( 5 ) achieved the highest yield (2.3 %) after 1 week at 35 C in the solid state of the six cataly sts that were trie d, although catalysts 8 and 9 did show some conversion as well. Any solid-state reaction that is occurr ing at this point is significant since these are presumably crystal line environment. At higher temperatures (50-55 C), once a certain yield is reached (410 %), a visible hete rogeneous change
PAGE 145
128 occurs, which we interpret as local melting poi nt depression from the presence of RCM product in the mixture, perhaps by a nuclea tion and growth mechanism about sites of defects in the crystal. This could indicate th at some amorphous character (in the form of defects) needs to be present in order for th is reaction to occur. This same phenomenon was observed with microwave experiments, where it was also shown that microwaves could accelerate RCM in the solid state, wh ich had previously been shown for liquid systems.
PAGE 146
129 LIST OF REFERENCES 1. Gavessotti, A.; Simonetta, M. Chem. Rev. 1982 82 1-13. 2. Tanaka, K.; Toda, F. Chem. Rev. 2000 100 1025-1074. 3. Mrinal, T. Encyclopedia of Polymer Science and Engineering ; Wiley: New York, N.Y., 1989; Vol. 15, p 362-380. 4. Pilati, F. Comprehensive Polymer Science ; Pergamon: New York, 1989; Vol. 5, p 201-216. 5. Herzberg, O.; Epple, M. Eur. J. Inorg. Chem. 2001 1395-1406. 6. Morawetz, H. J. Polym. Sci., Part C 1966 12 79-88. 7. Bamford, C. H.; Eastmond, G. C. Surface Defect Prop. Solids 1972 1 95-128. 8. Oakley, G. W.; Wagener, K. B. Molecula r catalysis and solid state polymerization. In Organotransition Metal Chemistry and Catalysis ; Marciniec, B., Ed.; Education In Advanced Chemistry, Vol. 8; Pozna -Wroc aw, 2002, 167-181. 9. Wegner, G. Faraday Discuss. R. Soc. Chem. 1979 68 494-508. 10. Bloor, D. Comprehensive Polymer Science ; Pergamon: New York, 1989; Vol. 5, p 233-249. 11. Baughman, R. H. J. Polym. Sci., Polym Phys. Ed. 1974 12 1511-1535. 12. Nishii, M.; Hayashi, K. Annu. Rev. Mater. Sci. 1975, 5 135-149. 13. Hasegawa, M. Chem. Rev. 1983 83 507-518. 14. Hasegawa, M. Comprehensive Polymer Science ; Pergamon: New York, 1989; Vol. 5, p 217-232. 15. Enkelmann, V. Acta. Cryst. B 1977 33 2842-2846. 16. Wegner, G. Makromol. Chem. 1971 145 85-94. 17. Enkelmann, V. Lect. Notes Phys. 1980 113 1-13. 18. Hasegawa, M.; Suzuki, Y. J. Polym. Sci., Part B 1967 5 813-815.
PAGE 147
130 19. Wegener, G. Mol. Cryst. Liq. Cryst. 1983 96 121-139. 20. Matsumoto, A. Prog. React. Kinet. Mech. 2001 26 59-110. 21. Morosoff, N.; Lim, D.; Morawetz, H. J. Am. Chem. Soc. 1964 86 3167. 22. Macchi, E. M.; Morosoff, N.; Morawetz, H. J. Polym. Sci., Part A-1 1968 6 2033-2049. 23. Chen, F. C.; Griskey, R. G.; Beyer, G. H. AIChE J. 1969 15 680-685. 24. Chang, T. M. Polym. Eng. Sci. 1970 10 364-368. 25. Gross, S. M.; Roberts, G. W.; Kiserow, D. J.; DeSimone, J. M. Macromolecules 2001 34 3916-3920. 26. Shi, C.; DeSimone, J. M.; Kiserow, D. J.; Roberts, G. W. Macromolecules 2001 34 7744-7750. 27. Gross, S. M.; Roberts, G. W.; Kiserow, D. J.; DeSimone, J. M. Macromolecules 2000 33 40-45. 28. Gross, S. M.; Flowers, D.; Roberts, G. W.; Kiserow, D. J.; DeSimone, J. M. Macromolecules 1999 32 3167-3169. 29. Macchi, E. M.; Morosoff, N.; Morawetz, H. J. Polym. Sci., Part A-1 1968 6 2033-2049. 30. Bamford, C. H.; Wayne, R. P. Polymer 1969 10 661-681. 31. Srinivasan, R.; Almonacil, C.; Naraya n, S.; Desai, P.; Abhiraman, A. S. Macromolecules 1998 31 6813-6821. 32. Ma, Y.; Agarwal, U. S.; Sikkema, D. J.; Lemstra, P. J. Polymer 2003 44 40854096. 33. Halpin, J. C. Polymer Letters 1964 2 955-963. 34. Li, L.; Huang, N.-X.; Liu, Z.-H .; Tang, Z.-L.; Yung, W.-S. Polym. Adv. Technol. 2000 11 242-249. 35. Goodner, M. D.; Gross, S. M.; DeSimone, J. M.; Roberts, G. W.; Kiserow, D. J. J. Appl. Polym. Sci. 2001 79 928-943. 36. Pilati, F. Comprehensive Polymer Science ; Pergamon: New York, 1989; Vol. 5, p 275-315. 37. Fitzgerald, J. A.; Irwin, R. S.; Memeger, W. Macromolecules 1991 24 32913299.
PAGE 148
131 38. Lenz, R. W.; Schuler, A. N. J. Polym. Sci., Polym. Symp. 1978 63 343-363. 39. Lenz, R. W.; Jin, J.-II, Feichtinger, K. A. Polymer 1983 24 327-334. 40. Handbook of Metathesis ; Grubbs, R. H., Ed.; Vol. 1; Wiley-VCH: Weinheim, 2003. 41. Handbook of Metathesis ; Grubbs, R. H., Ed.; Vol. 2; Wiley-VCH: Weinheim, 2003. 42. Handbook of Metathesis ; Grubbs, R. H., Ed.; Vol. 3; Wiley-VCH: Weinheim, 2003. 43. Alkene Metathesis in Organic Synthesis ; Frstner, A., Ed.; Topics in Organometallic Chemistry; Springer: Berlin, 1998. 44. Ivin, K. J.; Mol, J. C. Olefin Metathesis and Me tathesis Polymerization ; Academic Press: San Diego, CA, 1997. 45. Grubbs, R. H. Tetrahedron 2004 60 7117-7140. 46. Calderon, N. et al. Chem. Eng. News 1967 45 51. 47. Chatterjee, A. K. Olefin Cross-Metathesis. In Handbook of Metathesis ; Grubbs, R. H., Ed.; Wiley-VCH: Weinheim, 2003; Vol. 2, 246-295. 48. Randall, M. L.; Snapper, M. L. J. Mol. Catal. A: Chem. 1998 133 29-44. 49. Gibson, S. E.; Keen, S. P. Cross-Metathesis. In Alkene Metathesis in Organic Synthesis ; Frstner, A., Ed.; Topics in Organometallic Chemistry Series; Springer: Berlin, 1998, p 155-181. 50. Han, S.-Y.; Chang, S. General Ring-Closing Metathesis. In Handbook of Metathesis ; Grubbs, R. H., Ed.; Wiley-VC H: Weinheim, 2003; Vol. 2, 5-127. 51. Grubbs, R. H.; Miller, S. J.; Fu, G. C. Acc. Chem. Res ., 1995, 28 446-452. 52. Grubbs, R. H.; Chang, S. Tetrahedron 1998 54 4413-4450. 53. Hoveyda, A. H. Catalytic Ring-Closi ng Metathesis and the Development of Enantioselective Processes. In Alkene Metathesis in Organic Synthesis ; Frstner, A., Ed.; Topics in Organometallic Chem istry Series; Springer: Berlin, 1998, p 103132. 54. Hoveyda, A. H. Catalytic Asymmetric Olefin Metathesis. In Handbook of Metathesis ; Grubbs, R. H., Ed.; Wiley-VC H: Weinheim, 2003; Vol. 2, 128-152. 55. Schrader, T. O.; Snapper, M. L. Ring-Opening Cross-Metathesis. In Handbook of Metathesis ; Grubbs, R. H., Ed.; Wiley-VC H: Weinheim, 2003; Vol. 2, 205-237.
PAGE 149
132 56. Schrock, R. R. Acc. Chem. Res 1990 23 158-165. 57. Schrock, R. R. Ring Opening Metathesis Polymerization. In Ring Opening Polymerization ; Bruneile, D. J., Ed.; Hanser: Munich, 1993, p 129-156. 58. Grubbs, R. H.; Khosravi, E. Ring Open ing Metathesis Polymerization (ROMP) and Related Processess. In Synthesis of Polymers Schluter, A. D., Ed., Materials Science and Technology Series; Wiley: Weinheim, 1999, p 65-104. 59. Black, G.; Maher, D.; Risse, W. Li ving Ring-Opening Olefin Metathesis Polymerization. In Handbook of Metathesis ; Grubbs, R. H., Ed.; Wiley-VCH: Weinheim, 2003; Vol. 3, 2-71. 60. Lehman, S. E.; Wagener, K. B. ADMET Polymerization. In Handbook of Metathesis ; Grubbs, R. H., Ed.; Wiley-VC H: Weinheim, 2003; Vol. 3, 283-353. 61. Tindall, D.; Pawlow, J. H.; Wagener, K. B. Recent Advances in ADMET Chemistry. In Alkene Metathesis in Organic Synthesis ; Frstner, A., Ed.; Topics in Organometallic Chemistry Seri es; Springer: Berlin, 1998, p 183-198. 62. Davidson, T. A.; Wagener, K. B. Acyclic Diene Metathesis (ADMET) Polymerization. In Synthesis of Polymers Schluter, A. D., E d., Materials Science and Technology Serie; Wile y: Weinheim, 1999, p 105-122. 63. Wagener, K. B.; Valenti, D. Macromolecules 1997 30 6688-6690. 64. Smith, J. A.; Brzezinska, K. R.; Va lenti, D. J.; Wagener, K. B. Macromolecules 2000 33 3781-3794. 65. Sworen, J. C.; Smith, J. A.; Wagener, K. B.; Baugh, L. S.; Rucker, S. P. J. Am. Chem. Soc. 2003 125 2228-2240. 66. Truett, W. L.; Johnson, D. R.; R obinson, I. M.; Montague, B. A. J. Am. Chem. Soc. 1960 82 2337-2340. 67. Banks, R. L.; Bailey, G. C. Ind. Eng. Chem., Prod. Res. Dev. 1964 3 170-173. 68. Calderon, N.; Chen, H. Y.; Scott, K. W. Tetrahedron Lett. 1967 3327-3329. 69. Calderon, N.; Ofstead, E. A.; Judy, W. A. J. Polymer Sci., A-1 1967 5 2209-2217. 70. Bradshaw, C. P. C.; Howman, E. J.; Turner, L. J. Calalysis 1967 7 269-276. 71. Calderon, N.; Ofstead, E. A.; Ward, J. P.; Judy, W. A.; Scott, K, W. J. Am. Chem. Soc. 1968 90 4133-4140. 72. Hoffman, R.; Woodward, R. B. J. Am. Chem. Soc. 1965 87 2046-2048. 73. Mango, F. D.; Schachtschneider, J. H. J. Am. Chem. Soc. 1967 89 2484-2486.
PAGE 150
133 74. Calderon, N. Acc. Chem. Res. 1972 5 127-132. 75. Grubbs. R. H.; Brunck, T. K. J. Am. Chem. Soc. 1972 94 2538-2540. 76. Katz, T. J.; Cerefice, S. A. J. Am. Chem. Soc. 1969 91 6519-6521. 77. Cassar, L.; Eaton, P. E.; Halpern, J. J. Am. Chem. Soc. 1970 92 3515-3518. 78. Grubbs, R. H.; Burk, P. L.; Carr, D. D. J. Am. Chem. Soc. 1975 97 3265-3267. 79. Grubbs, R. H.; Carr, D. D.; Hoppin, C.; Burk, P. L. J. Am. Chem. Soc. 1976 98 3478-3483. 80. Herisson, J. L.; Chauvin, Y. Makromol. Chem. 1970 141 161-176. 81. Katz, T. J.; McGinnis, J. J. Am. Chem. Soc. 1975 97 ,1592-1594. 82. Schaverien, C. J.; Dewan, J. C.; Schrock, R. R. J. Am. Chem. Soc. 1986 108 2771-2773. 83. Murdzek, J. S.; Schrock, R. R. Organometallics 1987 6 1373-1374. 84. Schrock, R. R.; Hoveyda, A. H. Angew. Chem. Int. Ed. 2003 42 4592-4633. 85. Trnka, T. M.; Grubbs, R. H. Acc. Chem. Res. 2001 34 18-29. 86. Michelotti, F. W.; Keaveney, W. P. J. Polym. Sci. 1965 A3 895-905. 87. Rinehart, R. E.; Smith, H. P. Polym. Lett. 1965 3 1049-1052. 88. Novak, B. M.; Grubbs, R. H. J. Am. Chem. Soc. 1988 110 960-961. 89. Novak, B. M.; Grubbs, R. H. J. Am. Chem. Soc. 1988 110 7542-7543. 90. Lu, S.-Y.; Quayle, P.; Heatley, F.; Boot h, C.; Yeates, S. G.; Padget, J. C. Macromolecules 1992 25 2692-2697. 91. Nguyen, S. T.; Trnka, T. M. The Disc overy and Development of Well Defined, Ruthenium-Based Olefin Metathesis Catalysts. In Handbook of Metathesis ; Grubbs, R. H., Ed.; Wiley-VCH: Weinheim, 2003; Vol. 1, 61-85. 92. Nguyen, S. T.; Johnson, L. K.; Grubbs, R. H. J. Am. Chem. Soc. 1992 114 39743975. 93. Johnson, L. K.; Grubbs, R. H.; Ziller, J. W. J. Am. Chem. Soc. 1993 115 81308145. 94. Nguyen, S. T.; Grubbs, R. H. J. Am. Chem. Soc. 1993 115 9858-9859.
PAGE 151
134 95. Dias, E. L.; Nguyen, S. T.; Grubbs, R. H. J. Am. Chem. Soc. 1997 119 38873897. 96. Sanford, M. S.; Love, J. A.; Grubbs, R. H. J. Am. Chem. Soc. 2001 123 65436554. 97. Adlhart, C.; Chen, P. J. Am. Chem. Soc. 2004 126 3496-3510. 98. Schwab, P.; France, M. B.; Zill er, J. W.; Grubbs, R. H. Angew. Chem., Int. Ed 1995 34 2039-2041. 99. Schwab, P.; Grubbs, R. H.; Ziller, J. W. J. Am. Chem. Soc. 1996 118 100-110. 100. Wilhelm, T. E.; Berderrain, T. R.; Brown, S. N.; Grubbs, R. H. Organometallics 1997 16 3867-3869. 101. Belderrain, T. R.; Grubbs, R. H. Organometallics 1997 16 4001-4003. 102. Wolf, J.; Stuer, W.; Grunwald, C.; We rner, H.; Schwab, P.; Schulz, M. Angew. Chem., Int. Ed 1998 37 1124-1126. 103. Scholl, M.; Ding, S.; Lee, C. W.; Grubbs, R. H. Org. Lett 1999 1 953-956. 104. Scholl, M.; Trnka, T. M.; Morgan, J. P.; Grubbs, R. H. Tetrahedron Lett. 1999 40 2247-2250. 105. Bielawski, C. W.; Grubbs, R. H. Angew. Chem. Int. Ed. 2000 39 2903-2906. 106. Chatterjee, A. K.; Grubbs, R. H. Org. Lett. 1999 1 1751-1753. 107. Chatterjee, A. K.; Morgan, J. P.; Scholl, M.; Grubbs, R. H. J. Am. Chem. Soc. 2000 122 3783-3784. 108. Ulman, M.; Grubbs, R. H. J. Org. Chem. 1999 64 7202-7207. 109. Garber, S. B.; Kingsbury, J. S.; Grau, B. L.; Hoveyda, A. H. J. Am. Chem. Soc. 2000 122 8168-8179. 110. Frstner, A.; Ackerman, L.; Gabor, B.; Goddard, R.; Lehman, C. W.; Mynott, R.; Seltzer, F.; Thiel, O. R. Chem. Eur. J. 2001 7 3236-3253. 111. Dinger, M. B.; Mol, J. C. Adv. Synth. Catal. 2002 344 671-677. 112. Courchay, F. C.; Sworen, J. C.; Wagener, K. B. Macromolecules 2003 36 82318239. 113. Lehman, Jr., S. E.; Schwendeman, J. E. ; OÂ’Donnell, P. M.; Wagener, K. B. Inorg. Chim. Acta. 2003 345 190-198.
PAGE 152
135 114. Sworen, J. C.; Pawlow, J. H.; Case, W.; Lever, J.; Wagener, K. B. J. Mol. Catal. A: Chem. 2003 194 69-78. 115. Bourgeois, D.; Pancrazi, A.; Nolan, S. P.; Prunet, J. J Organomet. Chem. 2002 643-644 247-252. 116. Kinderman, S. S.; Maarseveen, J. H.; Schoemaker, H. E.; Hiemstra, H.; Rutjes, F. P. J. T. Org. Lett. 2001 3 2045-2048. 117. Thorn-Csanyi, E.; Dehmel, J.; Luginsland, H.–D.; Zilles, J. U. J. Mol. Catal. A: Chem. 1997 115 29-35. 118. McGrath, D. V.; Grubbs, R. H. Organometallics 1994 13 224-235. 119. Crabtree, R. H. The Organometallic Chemistry of the Transition Metals, 3rd Ed. Wiley: New York, 2001. p 226-228. 120. Hong, S. H.; Day, M. W.; Grubbs, R. H. J. Am. Chem. Soc. 2004 126 7414-7415. 121. Wagener, K. B.; Boncella, J. M.; Nel, J. G. Macromolecules 1991 24 2649-2657. 122. Sanford, M. S.; Ulman, M.; and Grubbs, R. H. J. Am. Chem. Soc. 2001 123 749750. 123. Louie, J.; Grubbs, R. H. Organometallics 2002 21 2153-2164. 124. Lindmark-Hamberg, M.; Wagener, K. B. Macromolecules 1987 20 2951-2954. 125. Wagener, K. B.; Boncella, J. M.; Nel, J. G.; Duttweiler, R. P.; Hillmyer, M. A. Makromol. Chem. 1990 191 365-374. 126. Wagener, K. B.; Nel, J. G.; Konzelman, J.; Boncella, J. M. Macromolecules 1990 23 5155-5157. 127. Sworen, J. C.; Smith, J. A.; Be rg, J. M.; Wagener, K. B. J. Am. Chem. Soc. 2004 126 11238-11246. 128. Wagener, K. B.; Smith, Jr., D. W. Macromolecules 1991 24 6073-6078. 129. Wagener, K. B.; Smith, Jr., D. W. Macromolecules 1993 26 1633-1642. 130. Wagener, K. B.; Smith, Jr., D. W. Macromolecules 1993 26, 3533-3537. 131. Brzezinska, K. R.; Schitter, R; Wagener, K. B. J. Polym. Sci: Part A 2000 38 1544-1550. 132. Church, A. C.; Pawlow, J. H.; Wagener, K. B. J. Organomet. Chem. 2001 620 287-295.
PAGE 153
136 133. Church, A. C.; Pawlow, J. H.; Wagener, K. B. Macromolecules 2002 35 57465751. 134. Church, A. C.; Pawlow, J. H.; Wagener, K. B. Macromol. Chem. Phys. 2003 204 32-39. 135. Wolfe, P. S.; Gomez, F. J.; Wagener, K. B. Macromolecules 1997 30 714-717. 136. Gomez, F. J.; Wagener, K. B. J. Organomet. Chem. 1999 592 271-277. 137. Hopkins, T. E.; Pawlow, J. H.; Koren, D. L.; Deters, K. S.; Solivan, S. M.; Davis, J. A.; Gomez, F. J.; Wagener, K. B. Macromolecules 2001 34 7920-7922. 138. Hopkins, T. E.; Wagener, K. B. Adv. Mater. 2002 14 1703-1715. 139. Hopkins, T. E.; Wagener, K. B. Macromolecules 2003 36 2206-2214. 140. Hopkins, T. E.; Wagener, K. B. Macromolecules 2004 37 1180-1189. 141. Hopkins, T. E.; Wagener, K. B. J. Mol. Catal. A: Chem. 2004 213 93-99. 142. OÂ’Donnell, P. M.; Brzezinska, K.; Powell, D.; Wagener, K. B. Macromolecules 2001 34 6845-6849. 143. OÂ’Donnell, P. M.; Wagener, K. B. J. Polym. Sci.: Part A 2003 41 2816-2827. 144. Wagener, K. B.; Brzezinska, K. Macromolecules 1991 24 5273-5277. 145. Wagener, K. B.; Brzezinska, K. Macromolecules 1992 25 2049-2052. 146. Patton, J. T.; Boncella, J. M.; Wagener, K. B. Macromolecules 1992 25 38623867. 147. Forbes, M. D. E.; Patton, J. T.; Myers, T. L.; Maynard, H. D.; Smith, Jr., D. W.; Schulz, G. R.; Wagener, K. B. J. Am. Chem. Soc. 1992 114 10978-10980. 148. Patton, J. T.; Wagener, K. B. Macromolecules 1993 26 249-253. 149. OÂ’Gara, J. E.; Portmess, J. D.; Wagener, K. B. Macromolecules 1993 26 28372841. 150. Tao, D.; Wagener, K. B. Macromolecules 1994 27 1281-1283. 151. Portmess, J. D.; Wagener, K. B. J. Polym. Sci.: Part A 1996 34 1353-1357. 152. Valenti, D. J.; Wagener, K. B. Macromolecules 1998 31 2764-2773. 153. Watson, M. D.; Wagener, K. B. Macromolecules 2000 33 3196-3201.
PAGE 154
137 154. Watson, M. D.; Wagener, K. B. Macromolecules 2000 33 5411-5417. 155. Watson, M. D.; Wagener, K. B. Macromolecules 2000 33 8963-8970. 156. Tindall, D.; Wagener, K. B. Macromolecules 2004 37 3328-3336. 157. Schwendeman, J. E.; Wagener, K. B. Macromolecules 2004 37 4031-4037. 158. Konzelman, J.; Wagener, K. B. Macromolecules 1995 28 4686-4692. 159. Wagener, K. B.; Brzezinska, K.; Anders on, J. D.; Younkin, T. R.; Steppe, K.; DeBoer, W. Macromolecules 1997 30 7363-7369. 160. Watson, M. D.; Wagener, K. B. Macromolecules 2000 33 1494-1496. 161. Lehman, Jr., S. E.; Wagener, K. B. Macromolecules 2002 35 48-53. 162. Matyjaszewski, K.; Xia, J. Chem. Rev. 2001 101 2921-2990. 163. Zantour, H.; Pousse, A; Mathilde, B. Bull. Soc. Chim. Fr. 1972 12 4715-4722. 164. Baughman, T. W.; Sworen, J. C.; Wagener, K. B. Tetrahedron 2004 60 1094310948. 165. Oakley, G. W.; Lehman, Jr., S. E.; Smith, J. A.; van Gerven, P.; Wagener, K. B. Macromolecules 2003 36 539-542. 166. Ravindranath, K; Mashelkar, R. A. Chem. Eng. Sci 1986 41 2197-2214. 167. Handbook of Conducting Polymers, Second Edition, Revised and Expanded ; Skotheim, T. A., Elsenbaumer, R. L., Re ynolds, J. R., Eds.; Marcel Dekker, Inc.: New York, 1998. 168. Reddinger, J. L.; Reynolds, J. R. Adv. Polym. Sci Wegner, G., Ed.; SpringerVerlag: Berlin, 1999, 145, 57-122. 169. Roncali, J. Chem. Rev. 1997 97 173-205. 170. Frommer, J. E.; Chance, R. R. Encyclopedia of Polymer Science and Engineering ; Wiley: New York, N.Y., 1989; Vol. 5, p 462-507. 171. Chandrasekhar, P. Conducting Polymers, Fundamentals and Applications ; Kluwer Academic Publishers: Boston, 1999. 172. Patil, A. O.; Heeger, A. J.; Wudl, F. Chem. Rev. 1988 88 183-200. 173. Tour, J. M. Chem. Rev. 1996 96 537-553. 174. Kavan, L. Chem. Rev. 1997 97 3061-3082.
PAGE 155
138 175. Shirakawa H.; Ikeda, S. Polymer J. 1971 2 231-244. 176. Ito, T.; Shirakawa, H.; Ikeda, S. J. Polym. Sci., Polym. Chem. Ed. 1974 12 11-20. 177. Shirakawa H.; Louis, E. J.; MacDiarmid, A. G.; Chiang, C. K.; Heeger, A. J. J. Chem. Soc., Chem. Commun. 1977 578-580. 178. Shirakawa, H Synth. Met. 2002 125 3-10. 179. MacDiarmid, A. G. Synth. Met. 2002 125 11-22. 180. Heeger, A. J. Synth. Met. 2002 125 23-42. 181. Wessling, R. A.; Zimmerman, R. G. Pol yxylylidenes and Method of Preparation. U. S. Patent 3,532,643, Oct. 6, 1970. 182. Wessling, R. A.; Zimmerman, R. G. Pol yxylylidenes Articles. U. S. Patent 3,706,677, Dec. 19, 1972. 183. Bradley, D. D. C.; Burn, P. L.; Friend, R. H.; Halliday, D. A.; Holmes, A. B.; Kraft, A. European Patent WO 93/14177, July 22, 1993. 184. Friend, R. H.; Gymer, R. W.; Holmes, A. B.; Burroughes, J. H.; Mraks, R. N.; Taliani, C.; Bradley, D. D. C.; Dos Sant os, D. A.; Bredas, J. L.; Logdlund, M.; Salaneck, W. R. Nature 1999 397 121-128. 185. Brabec, C. J.; Sariciftci, N. S.; Hummelen, J. C. Adv. Funct. Mater 2001 11 1526. 186. Peetz, R.; Strachota, A.; Thorn-Csnyi, E. Macromol. Chem. Phys. 2003 204, 1439-1450. 187. Thorn-Csnyi, E.; Kraxner, P.; Strachota, A. Macromol. Rapid Commun. 1998 19 223-228. 188. Thorn-Csnyi, E.; Kraxner, P. Macromol. Rapid Commun. 1995 16 147-153. 189. Irvin, J. A.; Reynolds, J. R. Polymer 1998 39 2339-2347. 190. Stalmach, U.; Detert, H. J. Prakt. Chem. 2000 34 10-16. 191. Thompson, B. C.; Madrigal, L. G.; Pint o, M.; Kang, T. –S.; Schanze, K. S.; Reynolds, J. R. "Donor-Acceptor Copolymers for Red and Near-Infrared Emitting PLEDs" J. Polym. Sci: Part A ., 2004 submitted 192. Oakley, G. W.; Wagener, K. B. Macromol. Chem. Phys. 2005 206 15-24. 193. Fu, G. C.; Nguyen, S. T.; Grubbs, R. H. J. Am. Chem. Soc. 1993 115 9856-9857.
PAGE 156
139 194. Mayo, K. G.; Nearhoof, E. H.; Kiddle, J. J. Org. Lett. 2002 4 1567-1570. 195. Thanh, G. V.; Loupy, A. Tetrahedron Lett. 2003 44 9091-9094. 196. Yang, C.; Murray, W. V.; Wilson, L. J. Tetrahedron Lett. 2003 44 1783-1786. 197. Efskind, J.; Undheim, K. Tetrahedron Lett. 2003 44 2837-2839. 198. Grigg, R.; Martin, W.; Morris, J.; Sridharan, V. Tetrahedron Lett. 2003 44 48994901. 199. Garbacia, S.; Desai, B.; Lavastre, O.; Kappe, C. O. J. Org. Chem. 2003 68 91369139. 200. For a review of microwave assisted organi c synthesis: Lidstrom, P.; Tierney, J.; Wathey, B.; Westman, J. Tetrahedron 2001 57 9225-9283.
PAGE 157
140 BIOGRAPHICAL SKETCH Garrett Wilmot Oakley was born in Southampton, NY, on December 7, 1977. At the age of three, his family moved to Calverton, NY where he lived until his graduation from Riverhead High School in 1996. Having w on the Rensselaer Polytechnic Institute Medal/Scholarship for his senior class, he pur sued a Bachelor of Science in chemistry at Rensselaer in Troy, NY, earning his degree in May 2000. While at Rensselaer, he became interested in polymer chemistry and conducted undergraduate research on reversible addition-fragmentation chain transf er (RAFT) agents under the supervision of Dr. Brian Benicewicz. Additiona lly, for the summer and fall semesters of his senior year, he did a co-op at the Bayer Corporation in Springfield, MA, where he worked in a research and development team working on c ontinuous mass ABS plastics. In June 2000, he moved to Gainesville, FL, where he began graduate study in organic/polymer chemistry at the University of Florida, in the group of Dr. Kenneth B. Wagener. While fundamental in scope, GarrettÂ’s dissertati on principally focuse d on conducting acyclic diene metathesis (ADMET) polycondensation in the solid state. On October 2, 2004, he married Jennifer Ann Sippel in Clearwater, FL, just over a month before the completion of his degree. As his father was the first in the Oakley family to receive a college education, Garrett was the first in the Oakley family to obtain a Doctoral degree when he received a Ph.D. in organic chemistry in December 2004.
|
